Block-Gibbs on Dirichlet Mixture Model¶
We will now see some of the key ingredients in action in a simple but more realistic setting and write a Dirichlet mixture model in GenJAX.
Clustering Points on the Real Line¶
The goal here is to cluster datapoints on the real line. To do so, we model a fixed number of clusters, each as a 1D-Gaussian with fixed variance, and we want to infer their means.
Model Description¶
The "model of the world" postulates:
- A fixed number of 1D Gaussians
- Each Gaussian is assigned a weight, representing the proportion of points assigned to each cluster
- Each datapoint is assigned to a cluster
Generative Process¶
We turn this into a generative model as follows:
- We have a fixed prior mean and variance for where the cluster centers might be
- We sample a mean for each cluster
- We sample an initial weight per cluster (sum of weights is 1)
- For each datapoint:
- We sample a cluster assignment proportional to the cluster weights
- We sample the datapoint noisily around the mean of the cluster
Implementation Details¶
We, the modelers, get to choose how this process is implemented.
- We choose distributions for each sampling step in a way that makes inference tractable.
- More precisely, we choose conjugate pairs so that we can do inference via Gibbs sampling.
- Gibbs sampling is an MCMC method that samples an initial trace, and then updates the traced choices we want to infer over time.
- To update a choice, Gibbs sampling samples from a conditional distribution, which is tractable with conjugate relationships.
In [1]:
Copied!
import genstudio.plot as Plot
import jax
import jax.numpy as jnp
import numpy as np
import genjax
from genjax import ChoiceMapBuilder as C
from genjax import categorical, dirichlet, gen, normal, pretty
from genjax._src.core.pytree import Const
pretty()
key = jax.random.key(0)
import genstudio.plot as Plot
import jax
import jax.numpy as jnp
import numpy as np
import genjax
from genjax import ChoiceMapBuilder as C
from genjax import categorical, dirichlet, gen, normal, pretty
from genjax._src.core.pytree import Const
pretty()
key = jax.random.key(0)
We define the generative model we described above. It has several hyperparameters that are somewhat manually inferred. An extension to the model could instead do inference over these hyperparameters, and fix hyper-hyperparameters instead.
In [2]:
Copied!
# Hyper parameters
PRIOR_VARIANCE = 10.0
OBS_VARIANCE = 1.0
N_DATAPOINTS = 5000
N_CLUSTERS = 40
ALPHA = float(N_DATAPOINTS / (N_CLUSTERS * 10))
PRIOR_MEAN = 50.0
N_ITER = 50
# Debugging mode
DEBUG = True
# Sub generative functions of the bigger model
@gen
def generate_cluster(mean, var):
cluster_mean = normal(mean, var) @ "mean"
return cluster_mean
@gen
def generate_cluster_weight(alphas):
probs = dirichlet(alphas) @ "probs"
return probs
@gen
def generate_datapoint(probs, clusters):
idx = categorical(jnp.log(probs)) @ "idx"
obs = normal(clusters[idx], OBS_VARIANCE) @ "obs"
return obs
@gen
def generate_datapoints(probs, clusters, n_datapoints):
idx = categorical(jnp.log(probs), sample_shape=n_datapoints) @ "idx"
obs = normal(clusters[idx], OBS_VARIANCE) @ "obs"
return obs
# Main model
@gen
def generate_data(n_clusters: Const[int], n_datapoints: Const[int], alpha: float):
clusters = (
generate_cluster.repeat(n=n_clusters.unwrap())(PRIOR_MEAN, PRIOR_VARIANCE)
@ "clusters"
)
probs = generate_cluster_weight.inline(
alpha / n_clusters.unwrap() * jnp.ones(n_clusters.unwrap())
)
datapoints = generate_datapoints(probs, clusters, n_datapoints) @ "datapoints"
return datapoints
# Hyper parameters
PRIOR_VARIANCE = 10.0
OBS_VARIANCE = 1.0
N_DATAPOINTS = 5000
N_CLUSTERS = 40
ALPHA = float(N_DATAPOINTS / (N_CLUSTERS * 10))
PRIOR_MEAN = 50.0
N_ITER = 50
# Debugging mode
DEBUG = True
# Sub generative functions of the bigger model
@gen
def generate_cluster(mean, var):
cluster_mean = normal(mean, var) @ "mean"
return cluster_mean
@gen
def generate_cluster_weight(alphas):
probs = dirichlet(alphas) @ "probs"
return probs
@gen
def generate_datapoint(probs, clusters):
idx = categorical(jnp.log(probs)) @ "idx"
obs = normal(clusters[idx], OBS_VARIANCE) @ "obs"
return obs
@gen
def generate_datapoints(probs, clusters, n_datapoints):
idx = categorical(jnp.log(probs), sample_shape=n_datapoints) @ "idx"
obs = normal(clusters[idx], OBS_VARIANCE) @ "obs"
return obs
# Main model
@gen
def generate_data(n_clusters: Const[int], n_datapoints: Const[int], alpha: float):
clusters = (
generate_cluster.repeat(n=n_clusters.unwrap())(PRIOR_MEAN, PRIOR_VARIANCE)
@ "clusters"
)
probs = generate_cluster_weight.inline(
alpha / n_clusters.unwrap() * jnp.ones(n_clusters.unwrap())
)
datapoints = generate_datapoints(probs, clusters, n_datapoints) @ "datapoints"
return datapoints
We create some synthetic data to test inference.
In [3]:
Copied!
# Generate synthetic data with N_CLUSTERS clusters evenly spaced
points_per_cluster = int(N_DATAPOINTS / N_CLUSTERS)
cluster_indices = jnp.arange(N_CLUSTERS)
offsets = PRIOR_VARIANCE * (-4 + 8 * cluster_indices / N_CLUSTERS)
# Create keys for each cluster
keys = jax.random.split(jax.random.key(0), N_CLUSTERS)
# Generate uniform random points for each cluster
uniform_points = jax.vmap(lambda k: jax.random.uniform(k, shape=(points_per_cluster,)))(
keys
)
# Add offset and prior mean to each cluster's points
shifted_points = uniform_points + (PRIOR_MEAN + offsets[:, None])
datapoints = C["datapoints", "obs"].set(shifted_points.reshape(-1))
# Generate synthetic data with N_CLUSTERS clusters evenly spaced
points_per_cluster = int(N_DATAPOINTS / N_CLUSTERS)
cluster_indices = jnp.arange(N_CLUSTERS)
offsets = PRIOR_VARIANCE * (-4 + 8 * cluster_indices / N_CLUSTERS)
# Create keys for each cluster
keys = jax.random.split(jax.random.key(0), N_CLUSTERS)
# Generate uniform random points for each cluster
uniform_points = jax.vmap(lambda k: jax.random.uniform(k, shape=(points_per_cluster,)))(
keys
)
# Add offset and prior mean to each cluster's points
shifted_points = uniform_points + (PRIOR_MEAN + offsets[:, None])
datapoints = C["datapoints", "obs"].set(shifted_points.reshape(-1))
We now write the main inference loop. As we said at the beginning, we do MCMC via Gibbs sampling. Inference therefore consist of a main loop and we evolve a trace over time. The final trace contains a sample from the approximate posterior.
In [4]:
Copied!
def infer(datapoints):
key = jax.random.key(32421)
args = (Const(N_CLUSTERS), Const(N_DATAPOINTS), ALPHA)
key, subkey = jax.random.split(key)
initial_weights = C["probs"].set(jnp.ones(N_CLUSTERS) / N_CLUSTERS)
constraints = datapoints | initial_weights
tr, _ = generate_data.importance(subkey, constraints, args)
if DEBUG:
all_posterior_means = [tr.get_choices()["clusters", "mean"]]
all_posterior_weights = [tr.get_choices()["probs"]]
all_cluster_assignment = [tr.get_choices()["datapoints", "idx"]]
jax.debug.print("Initial means: {v}", v=all_posterior_means[0])
jax.debug.print("Initial weights: {v}", v=all_posterior_weights[0])
for _ in range(N_ITER):
# Gibbs update on `("clusters", i, "mean")` for each i, in parallel
key, subkey = jax.random.split(key)
tr = jax.jit(update_cluster_means)(subkey, tr)
all_posterior_means.append(tr.get_choices()["clusters", "mean"])
# # Gibbs update on `("datapoints", i, "idx")` for each `i`, in parallel
key, subkey = jax.random.split(key)
tr = jax.jit(update_datapoint_assignment)(subkey, tr)
all_cluster_assignment.append(tr.get_choices()["datapoints", "idx"])
# # Gibbs update on `probs`
key, subkey = jax.random.split(key)
tr = jax.jit(update_cluster_weights)(subkey, tr)
all_posterior_weights.append(tr.get_choices()["probs"])
return all_posterior_means, all_posterior_weights, all_cluster_assignment, tr
else:
# One Gibbs sweep consist of updating each latent variable
def update(carry, _):
key, tr = carry
# Gibbs update on `("clusters", i, "mean")` for each i, in parallel
key, subkey = jax.random.split(key)
tr = update_cluster_means(subkey, tr)
# Gibbs update on `("datapoints", i, "idx")` for each `i`, in parallel
key, subkey = jax.random.split(key)
tr = update_datapoint_assignment(subkey, tr)
# Gibbs update on `probs`
key, subkey = jax.random.split(key)
tr = update_cluster_weights(subkey, tr)
return (key, tr), None
# Overall inference performs a fixed number of Gibbs sweeps
(key, tr), _ = jax.jit(jax.lax.scan)(update, (key, tr), None, length=N_ITER)
return tr
def update_cluster_means(key, trace):
# We can update each cluster in parallel
# For each cluster, we find the datapoints in that cluster and compute their mean
datapoint_indexes = trace.get_choices()["datapoints", "idx"]
datapoints = trace.get_choices()["datapoints", "obs"]
n_clusters = trace.get_args()[0].unwrap()
current_means = trace.get_choices()["clusters", "mean"]
# Count number of points per cluster
category_counts = jnp.bincount(
trace.get_choices()["datapoints", "idx"],
length=n_clusters,
minlength=n_clusters,
)
# Will contain some NaN due to clusters having no datapoint
cluster_means = (
jax.vmap(
lambda i: jnp.sum(jnp.where(datapoint_indexes == i, datapoints, 0)),
in_axes=(0),
out_axes=(0),
)(jnp.arange(n_clusters))
/ category_counts
)
# Conjugate update for Normal-iid-Normal distribution
# See https://people.eecs.berkeley.edu/~jordan/courses/260-spring10/lectures/lecture5.pdf
# Note that there's a typo in the math for the posterior mean.
posterior_means = (
PRIOR_VARIANCE
/ (PRIOR_VARIANCE + OBS_VARIANCE / category_counts)
* cluster_means
+ (OBS_VARIANCE / category_counts)
/ (PRIOR_VARIANCE + OBS_VARIANCE / category_counts)
* PRIOR_MEAN
)
posterior_variances = 1 / (1 / PRIOR_VARIANCE + category_counts / OBS_VARIANCE)
# Gibbs resampling of cluster means
key, subkey = jax.random.split(key)
new_means = (
generate_cluster.vmap()
.simulate(key, (posterior_means, posterior_variances))
.get_choices()["mean"]
)
# Remove the sampled Nan due to clusters having no datapoint and pick previous mean in that case, i.e. no Gibbs update for them
chosen_means = jnp.where(category_counts == 0, current_means, new_means)
if DEBUG:
jax.debug.print("Category counts: {v}", v=category_counts)
jax.debug.print("Current means: {v}", v=cluster_means)
jax.debug.print("Posterior means: {v}", v=posterior_means)
jax.debug.print(fmt="Posterior variance: {v}", v=posterior_variances)
jax.debug.print("Resampled means: {v}", v=new_means)
jax.debug.print("Chosen means: {v}", v=chosen_means)
argdiffs = genjax.Diff.no_change(trace.args)
new_trace, _, _, _ = trace.update(
subkey, C["clusters", "mean"].set(chosen_means), argdiffs
)
return new_trace
def update_datapoint_assignment(key, trace):
# We want to update the index for each datapoint, in parallel.
# It means we want to resample the i, but instead of being from the prior
# P(i | probs), we do it from the local posterior P(i | probs, xs).
# We need to do it for all addresses ["datapoints", "idx", i],
# and as these are independent (when conditioned on the rest)
# we can resample them in parallel.
# Conjugate update for a categorical is just exact posterior via enumeration
# P(x | y ) = P(x, y) \ sum_x P(x, y).
# P(x | y1, y2) = P(x | y1)
# Sampling from
# (P(x = 1 | y ), P(x = 2 | y), ...) is the same as
# sampling from Categorical(P(x = 1, y), P(x = 2, y))
# as the weights need not be normalized
# In addition, if the model factorizes as P(x, y1, y2) = P(x, y1)P(y1 | y2),
# we can further simplify P(y1 | y2) from the categorical as it does not depend on x. More generally We only need to look at the children and parents of x ("idx" in our situation, which are conveniently wrapped in the generate_datapoint generative function).
def compute_local_density(x, i):
datapoint_mean = trace.get_choices()["datapoints", "obs", x]
chm = C["obs"].set(datapoint_mean).at["idx"].set(i)
clusters = trace.get_choices()["clusters", "mean"]
probs = trace.get_choices()["probs"]
args = (probs, clusters)
model_logpdf, _ = generate_datapoint.assess(chm, args)
return model_logpdf
n_clusters = trace.get_args()[0].unwrap()
n_datapoints = trace.get_args()[1].unwrap()
local_densities = jax.vmap(
lambda x: jax.vmap(lambda i: compute_local_density(x, i))(
jnp.arange(n_clusters)
)
)(jnp.arange(n_datapoints))
# Conjugate update by sampling from posterior categorical
# Note: I think we could've used something like
# generate_datapoint.vmap().importance which would perhaps
# work in a more general setting but would definitely be slower here.
key, subkey = jax.random.split(key)
new_datapoint_indexes = genjax.categorical.simulate(
key, (local_densities,)
).get_choices()
# Gibbs resampling of datapoint assignment to clusters
argdiffs = genjax.Diff.no_change(trace.args)
new_trace, _, _, _ = trace.update(
subkey, C["datapoints", "idx"].set(new_datapoint_indexes), argdiffs
)
return new_trace
def update_cluster_weights(key, trace):
# Count number of points per cluster
n_clusters = trace.get_args()[0].unwrap()
category_counts = jnp.bincount(
trace.get_choices()["datapoints", "idx"],
length=n_clusters,
minlength=n_clusters,
)
# Conjugate update for Dirichlet distribution
# See https://en.wikipedia.org/wiki/Dirichlet_distribution#Conjugate_to_categorical_or_multinomial
new_alpha = ALPHA / n_clusters * jnp.ones(n_clusters) + category_counts
# Gibbs resampling of cluster weights
key, subkey = jax.random.split(key)
new_probs = generate_cluster_weight.simulate(key, (new_alpha,)).get_retval()
if DEBUG:
jax.debug.print(fmt="Category counts: {v}", v=category_counts)
jax.debug.print(fmt="New alpha: {v}", v=new_alpha)
jax.debug.print(fmt="New probs: {v}", v=new_probs)
argdiffs = genjax.Diff.no_change(trace.args)
new_trace, _, _, _ = trace.update(subkey, C["probs"].set(new_probs), argdiffs)
return new_trace
def infer(datapoints):
key = jax.random.key(32421)
args = (Const(N_CLUSTERS), Const(N_DATAPOINTS), ALPHA)
key, subkey = jax.random.split(key)
initial_weights = C["probs"].set(jnp.ones(N_CLUSTERS) / N_CLUSTERS)
constraints = datapoints | initial_weights
tr, _ = generate_data.importance(subkey, constraints, args)
if DEBUG:
all_posterior_means = [tr.get_choices()["clusters", "mean"]]
all_posterior_weights = [tr.get_choices()["probs"]]
all_cluster_assignment = [tr.get_choices()["datapoints", "idx"]]
jax.debug.print("Initial means: {v}", v=all_posterior_means[0])
jax.debug.print("Initial weights: {v}", v=all_posterior_weights[0])
for _ in range(N_ITER):
# Gibbs update on `("clusters", i, "mean")` for each i, in parallel
key, subkey = jax.random.split(key)
tr = jax.jit(update_cluster_means)(subkey, tr)
all_posterior_means.append(tr.get_choices()["clusters", "mean"])
# # Gibbs update on `("datapoints", i, "idx")` for each `i`, in parallel
key, subkey = jax.random.split(key)
tr = jax.jit(update_datapoint_assignment)(subkey, tr)
all_cluster_assignment.append(tr.get_choices()["datapoints", "idx"])
# # Gibbs update on `probs`
key, subkey = jax.random.split(key)
tr = jax.jit(update_cluster_weights)(subkey, tr)
all_posterior_weights.append(tr.get_choices()["probs"])
return all_posterior_means, all_posterior_weights, all_cluster_assignment, tr
else:
# One Gibbs sweep consist of updating each latent variable
def update(carry, _):
key, tr = carry
# Gibbs update on `("clusters", i, "mean")` for each i, in parallel
key, subkey = jax.random.split(key)
tr = update_cluster_means(subkey, tr)
# Gibbs update on `("datapoints", i, "idx")` for each `i`, in parallel
key, subkey = jax.random.split(key)
tr = update_datapoint_assignment(subkey, tr)
# Gibbs update on `probs`
key, subkey = jax.random.split(key)
tr = update_cluster_weights(subkey, tr)
return (key, tr), None
# Overall inference performs a fixed number of Gibbs sweeps
(key, tr), _ = jax.jit(jax.lax.scan)(update, (key, tr), None, length=N_ITER)
return tr
def update_cluster_means(key, trace):
# We can update each cluster in parallel
# For each cluster, we find the datapoints in that cluster and compute their mean
datapoint_indexes = trace.get_choices()["datapoints", "idx"]
datapoints = trace.get_choices()["datapoints", "obs"]
n_clusters = trace.get_args()[0].unwrap()
current_means = trace.get_choices()["clusters", "mean"]
# Count number of points per cluster
category_counts = jnp.bincount(
trace.get_choices()["datapoints", "idx"],
length=n_clusters,
minlength=n_clusters,
)
# Will contain some NaN due to clusters having no datapoint
cluster_means = (
jax.vmap(
lambda i: jnp.sum(jnp.where(datapoint_indexes == i, datapoints, 0)),
in_axes=(0),
out_axes=(0),
)(jnp.arange(n_clusters))
/ category_counts
)
# Conjugate update for Normal-iid-Normal distribution
# See https://people.eecs.berkeley.edu/~jordan/courses/260-spring10/lectures/lecture5.pdf
# Note that there's a typo in the math for the posterior mean.
posterior_means = (
PRIOR_VARIANCE
/ (PRIOR_VARIANCE + OBS_VARIANCE / category_counts)
* cluster_means
+ (OBS_VARIANCE / category_counts)
/ (PRIOR_VARIANCE + OBS_VARIANCE / category_counts)
* PRIOR_MEAN
)
posterior_variances = 1 / (1 / PRIOR_VARIANCE + category_counts / OBS_VARIANCE)
# Gibbs resampling of cluster means
key, subkey = jax.random.split(key)
new_means = (
generate_cluster.vmap()
.simulate(key, (posterior_means, posterior_variances))
.get_choices()["mean"]
)
# Remove the sampled Nan due to clusters having no datapoint and pick previous mean in that case, i.e. no Gibbs update for them
chosen_means = jnp.where(category_counts == 0, current_means, new_means)
if DEBUG:
jax.debug.print("Category counts: {v}", v=category_counts)
jax.debug.print("Current means: {v}", v=cluster_means)
jax.debug.print("Posterior means: {v}", v=posterior_means)
jax.debug.print(fmt="Posterior variance: {v}", v=posterior_variances)
jax.debug.print("Resampled means: {v}", v=new_means)
jax.debug.print("Chosen means: {v}", v=chosen_means)
argdiffs = genjax.Diff.no_change(trace.args)
new_trace, _, _, _ = trace.update(
subkey, C["clusters", "mean"].set(chosen_means), argdiffs
)
return new_trace
def update_datapoint_assignment(key, trace):
# We want to update the index for each datapoint, in parallel.
# It means we want to resample the i, but instead of being from the prior
# P(i | probs), we do it from the local posterior P(i | probs, xs).
# We need to do it for all addresses ["datapoints", "idx", i],
# and as these are independent (when conditioned on the rest)
# we can resample them in parallel.
# Conjugate update for a categorical is just exact posterior via enumeration
# P(x | y ) = P(x, y) \ sum_x P(x, y).
# P(x | y1, y2) = P(x | y1)
# Sampling from
# (P(x = 1 | y ), P(x = 2 | y), ...) is the same as
# sampling from Categorical(P(x = 1, y), P(x = 2, y))
# as the weights need not be normalized
# In addition, if the model factorizes as P(x, y1, y2) = P(x, y1)P(y1 | y2),
# we can further simplify P(y1 | y2) from the categorical as it does not depend on x. More generally We only need to look at the children and parents of x ("idx" in our situation, which are conveniently wrapped in the generate_datapoint generative function).
def compute_local_density(x, i):
datapoint_mean = trace.get_choices()["datapoints", "obs", x]
chm = C["obs"].set(datapoint_mean).at["idx"].set(i)
clusters = trace.get_choices()["clusters", "mean"]
probs = trace.get_choices()["probs"]
args = (probs, clusters)
model_logpdf, _ = generate_datapoint.assess(chm, args)
return model_logpdf
n_clusters = trace.get_args()[0].unwrap()
n_datapoints = trace.get_args()[1].unwrap()
local_densities = jax.vmap(
lambda x: jax.vmap(lambda i: compute_local_density(x, i))(
jnp.arange(n_clusters)
)
)(jnp.arange(n_datapoints))
# Conjugate update by sampling from posterior categorical
# Note: I think we could've used something like
# generate_datapoint.vmap().importance which would perhaps
# work in a more general setting but would definitely be slower here.
key, subkey = jax.random.split(key)
new_datapoint_indexes = genjax.categorical.simulate(
key, (local_densities,)
).get_choices()
# Gibbs resampling of datapoint assignment to clusters
argdiffs = genjax.Diff.no_change(trace.args)
new_trace, _, _, _ = trace.update(
subkey, C["datapoints", "idx"].set(new_datapoint_indexes), argdiffs
)
return new_trace
def update_cluster_weights(key, trace):
# Count number of points per cluster
n_clusters = trace.get_args()[0].unwrap()
category_counts = jnp.bincount(
trace.get_choices()["datapoints", "idx"],
length=n_clusters,
minlength=n_clusters,
)
# Conjugate update for Dirichlet distribution
# See https://en.wikipedia.org/wiki/Dirichlet_distribution#Conjugate_to_categorical_or_multinomial
new_alpha = ALPHA / n_clusters * jnp.ones(n_clusters) + category_counts
# Gibbs resampling of cluster weights
key, subkey = jax.random.split(key)
new_probs = generate_cluster_weight.simulate(key, (new_alpha,)).get_retval()
if DEBUG:
jax.debug.print(fmt="Category counts: {v}", v=category_counts)
jax.debug.print(fmt="New alpha: {v}", v=new_alpha)
jax.debug.print(fmt="New probs: {v}", v=new_probs)
argdiffs = genjax.Diff.no_change(trace.args)
new_trace, _, _, _ = trace.update(subkey, C["probs"].set(new_probs), argdiffs)
return new_trace
We can now run inference, obtaining the final trace and some intermediate traces for visualizing inference.
In [5]:
Copied!
if DEBUG:
(
all_posterior_means,
all_posterior_weights,
all_cluster_assignment,
posterior_trace,
) = infer(datapoints)
else:
posterior_trace = infer(datapoints)
posterior_trace
if DEBUG:
(
all_posterior_means,
all_posterior_weights,
all_cluster_assignment,
posterior_trace,
) = infer(datapoints)
else:
posterior_trace = infer(datapoints)
posterior_trace
Initial means: [40.161377 32.067596 43.754623 58.22524 71.114914 48.64424 42.166443 49.68216 58.838997 45.826347 67.565094 54.852512 46.537666 61.92414 47.47585 37.81533 58.96495 54.572662 53.413666 57.45804 50.388985 47.83156 47.369118 60.489502 38.442974 57.56394 61.01972 58.002117 63.748646 51.57735 38.8523 61.786335 57.70788 46.753838 62.71544 65.316536 53.039482 62.816944 44.667416 52.880417] Initial weights: [0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025 0.025]
Category counts: [112 138 113 112 134 132 109 125 137 111 115 131 111 114 120 150 125 133 135 124 118 108 134 123 136 122 119 133 108 126 116 127 143 153 126 119 111 127 120 150] Current means: [48.551105 50.543533 49.47744 47.197395 53.045887 50.91979 50.41066 48.702236 48.789185 53.10006 48.135967 50.173283 53.305874 50.963715 48.833 46.289837 48.820015 48.592518 50.06069 43.21971 49.662766 51.40882 49.092323 52.790802 49.188526 51.837326 48.10093 49.705772 45.944798 49.341793 47.11504 49.200302 48.26917 49.685734 49.80089 49.685 50.09544 48.22027 50.925735 51.197903] Posterior variance: [0.00892061 0.00724113 0.00884173 0.00892061 0.00745712 0.00757002 0.0091659 0.0079936 0.00729395 0.0090009 0.0086881 0.00762776 0.0090009 0.00876424 0.00832639 0.00666222 0.0079936 0.00751315 0.00740192 0.00805802 0.0084674 0.00925069 0.00745712 0.00812348 0.00734754 0.00819001 0.00839631 0.00751315 0.00925069 0.00793021 0.00861326 0.00786782 0.00698812 0.00653168 0.00793021 0.00839631 0.0090009 0.00786782 0.00832639 0.00666222] Posterior means: [48.5524 50.54314 49.477905 47.199894 53.043613 50.91909 50.410282 48.70327 48.790066 53.097267 48.13759 50.173153 53.302895 50.962868 48.833973 46.292305 48.820953 48.59358 50.06065 43.225178 49.663048 51.407516 49.093 52.788536 49.189125 51.83582 48.102524 49.705994 45.948547 49.34231 47.117527 49.20093 48.270374 49.68594 49.80105 49.685265 50.095352 48.22167 50.92497 51.1971 ] Resampled means: [48.55825 50.55419 49.470776 47.21506 53.04305 50.919296 50.41241 48.702618 48.78646 53.088394 48.130272 50.174213 53.30334 50.96362 48.851543 46.293194 48.818577 48.60128 50.05313 43.226105 49.644917 51.40527 49.08746 52.78132 49.196877 51.830276 48.12212 49.710495 45.938034 49.34437 47.11272 49.205276 48.27407 49.67587 49.80486 49.684532 50.112442 48.228474 50.92309 51.20577 ] Chosen means: [48.55825 50.55419 49.470776 47.21506 53.04305 50.919296 50.41241 48.702618 48.78646 53.088394 48.130272 50.174213 53.30334 50.96362 48.851543 46.293194 48.818577 48.60128 50.05313 43.226105 49.644917 51.40527 49.08746 52.78132 49.196877 51.830276 48.12212 49.710495 45.938034 49.34437 47.11272 49.205276 48.27407 49.67587 49.80486 49.684532 50.112442 48.228474 50.92309 51.20577 ]
Category counts: [ 9 17 7 15 122 13 8 11 13 155 11 12 1980 13
9 56 9 11 12 2180 5 11 9 51 7 20 16 10
66 7 28 12 14 8 10 10 9 10 18 16]
New alpha: [ 9.3125 17.3125 7.3125 15.3125 122.3125 13.3125 8.3125
11.3125 13.3125 155.3125 11.3125 12.3125 1980.3125 13.3125
9.3125 56.3125 9.3125 11.3125 12.3125 2180.3125 5.3125
11.3125 9.3125 51.3125 7.3125 20.3125 16.3125 10.3125
66.3125 7.3125 28.3125 12.3125 14.3125 8.3125 10.3125
10.3125 9.3125 10.3125 18.3125 16.3125]
New probs: [0.00189926 0.00248475 0.0011278 0.00288892 0.0236311 0.00198729
0.00118441 0.00294586 0.00234889 0.03112511 0.0016432 0.00269484
0.40460277 0.00145802 0.00170202 0.01265933 0.00196078 0.00179546
0.00420049 0.43071568 0.00069633 0.00147408 0.00237535 0.01011323
0.00069506 0.00471409 0.00409952 0.00223707 0.01044026 0.00110905
0.00644143 0.0016642 0.00338145 0.00168727 0.00287666 0.00158927
0.00199235 0.00114861 0.00308817 0.00312046]
Category counts: [ 9 17 7 15 122 13 8 11 13 155 11 12 1980 13
9 56 9 11 12 2180 5 11 9 51 7 20 16 10
66 7 28 12 14 8 10 10 9 10 18 16]
Current means: [48.3503 51.0182 49.699253 46.7219 57.953323 51.28109 51.09111
48.378513 48.29023 58.85107 48.412537 50.050426 73.18892 51.23476
48.858047 45.76423 48.422256 48.53585 50.32243 26.954638 49.733826
52.249588 49.431244 54.54483 49.46888 52.63449 47.459064 49.56484
45.256035 49.20731 46.395695 48.940018 47.703197 50.104774 49.71475
49.669403 50.19495 47.578484 51.26888 51.89612 ]
Posterior variance: [0.1098901 0.05847953 0.14084508 0.06622516 0.00819001 0.07633588
0.12345678 0.09009009 0.07633588 0.00644745 0.09009009 0.08264463
0.00050503 0.07633588 0.1098901 0.01782531 0.1098901 0.09009009
0.08264463 0.00045869 0.19607843 0.09009009 0.1098901 0.01956947
0.14084508 0.04975124 0.0621118 0.09900989 0.01512859 0.14084508
0.03558719 0.08264463 0.07092199 0.12345678 0.09900989 0.09900989
0.1098901 0.09900989 0.05524862 0.0621118 ]
Posterior means: [48.36843 51.012245 49.703487 46.743614 57.946808 51.27131 51.07764
48.393124 48.30328 58.845364 48.426838 50.05001 73.187744 51.225334
48.870598 45.77178 48.439598 48.54904 50.319767 26.955692 49.739044
52.22932 49.437496 54.53594 49.476357 52.621384 47.474846 49.56914
45.26321 49.21847 46.408524 48.94878 47.71949 50.10348 49.717567
49.672672 50.19281 47.602455 51.261868 51.884342]
Resampled means: [48.26658 51.035984 49.44021 46.686794 57.94227 51.224907 51.103603
48.445908 48.296204 58.846733 48.42595 50.088627 73.18736 51.085728
48.872852 45.76648 48.36272 48.377502 50.316067 26.956404 49.55857
52.278164 49.283646 54.53155 49.407547 52.575245 47.37057 49.63649
45.261692 49.10549 46.400024 48.991364 47.75156 50.06408 49.37957
49.885742 50.077755 47.58485 51.334045 51.865147]
Chosen means: [48.26658 51.035984 49.44021 46.686794 57.94227 51.224907 51.103603
48.445908 48.296204 58.846733 48.42595 50.088627 73.18736 51.085728
48.872852 45.76648 48.36272 48.377502 50.316067 26.956404 49.55857
52.278164 49.283646 54.53155 49.407547 52.575245 47.37057 49.63649
45.261692 49.10549 46.400024 48.991364 47.75156 50.06408 49.37957
49.885742 50.077755 47.58485 51.334045 51.865147]
Category counts: [ 6 13 4 17 154 13 12 17 13 454 9 16 1499 15
7 138 11 10 29 1661 0 9 12 142 2 56 25 8
501 4 30 3 12 8 15 7 8 6 19 35]
New alpha: [6.3125000e+00 1.3312500e+01 4.3125000e+00 1.7312500e+01 1.5431250e+02
1.3312500e+01 1.2312500e+01 1.7312500e+01 1.3312500e+01 4.5431250e+02
9.3125000e+00 1.6312500e+01 1.4993125e+03 1.5312500e+01 7.3125000e+00
1.3831250e+02 1.1312500e+01 1.0312500e+01 2.9312500e+01 1.6613125e+03
3.1250000e-01 9.3125000e+00 1.2312500e+01 1.4231250e+02 2.3125000e+00
5.6312500e+01 2.5312500e+01 8.3125000e+00 5.0131250e+02 4.3125000e+00
3.0312500e+01 3.3125000e+00 1.2312500e+01 8.3125000e+00 1.5312500e+01
7.3125000e+00 8.3125000e+00 6.3125000e+00 1.9312500e+01 3.5312500e+01]
New probs: [1.7880501e-03 2.6458497e-03 6.5198482e-04 3.2564767e-03 2.9793091e-02
2.4859312e-03 2.1520744e-03 5.6812679e-03 2.1166757e-03 9.0967804e-02
2.7477716e-03 3.4865218e-03 2.9469055e-01 2.6578207e-03 1.3730587e-03
2.6505841e-02 2.9044235e-03 2.1698035e-03 4.5546470e-03 3.3812034e-01
1.4779117e-05 1.9852251e-03 1.5465210e-03 2.8276537e-02 2.2804660e-04
1.1101911e-02 3.8824102e-03 2.7555102e-03 1.0261990e-01 5.8655109e-04
7.2166594e-03 4.6558920e-04 2.3459627e-03 9.7187527e-04 1.6815295e-03
1.5860586e-03 1.0851868e-03 2.1350540e-03 4.3881638e-03 4.3765390e-03]
Category counts: [ 6 13 4 17 154 13 12 17 13 454 9 16 1499 15
7 138 11 10 29 1661 0 9 12 142 2 56 25 8
501 4 30 3 12 8 15 7 8 6 19 35]
Current means: [47.81455 51.616352 49.277615 47.094162 57.616882 51.725407 51.252853
48.244137 48.34731 61.638596 48.172947 50.797592 77.4957 51.253555
49.20161 43.85784 48.182377 47.83187 50.30712 22.795555 nan
52.190456 49.17903 54.544834 49.06475 52.970554 47.385933 49.942406
40.528355 48.83827 46.231007 49.321194 47.58856 50.83484 49.565285
50.272373 50.259727 47.38881 51.33322 51.805378]
Posterior variance: [1.6393442e-01 7.6335877e-02 2.4390244e-01 5.8479533e-02 6.4892923e-03
7.6335877e-02 8.2644626e-02 5.8479533e-02 7.6335877e-02 2.2021581e-03
1.0989010e-01 6.2111799e-02 6.6706690e-04 6.6225164e-02 1.4084508e-01
7.2411294e-03 9.0090089e-02 9.9009894e-02 3.4364261e-02 6.0201070e-04
1.0000000e+01 1.0989010e-01 8.2644626e-02 7.0372974e-03 4.7619051e-01
1.7825313e-02 3.9840639e-02 1.2345678e-01 1.9956096e-03 2.4390244e-01
3.3222590e-02 3.2258067e-01 8.2644626e-02 1.2345678e-01 6.6225164e-02
1.4084508e-01 1.2345678e-01 1.6393442e-01 5.2356020e-02 2.8490029e-02]
Posterior means: [47.850372 51.60401 49.295235 47.111156 57.611942 51.712234 51.2425
48.254406 48.359924 61.63603 48.193027 50.792637 77.49386 51.245255
49.212852 43.86229 48.198753 47.853333 50.306065 22.797194 nan
52.16639 49.185818 54.541637 49.109283 52.965256 47.39635 49.943115
40.530247 48.866604 46.243526 49.343094 47.60849 50.82453 49.568165
50.268536 50.25652 47.431614 51.326244 51.800232]
Resampled means: [47.571293 51.561016 49.57103 47.065582 57.60884 51.71842 51.147987
48.30018 48.338913 61.638508 48.23913 50.86512 77.4943 51.187626
49.08549 43.863163 48.39046 47.856056 50.29154 22.796968 nan
52.2755 49.18443 54.541786 49.078922 52.96913 47.357506 49.91127
40.52969 48.609577 46.219105 48.567707 47.50982 50.94427 49.646065
50.107346 50.30451 47.274677 51.30897 51.84024 ]
Chosen means: [47.571293 51.561016 49.57103 47.065582 57.60884 51.71842 51.147987
48.30018 48.338913 61.638508 48.23913 50.86512 77.4943 51.187626
49.08549 43.863163 48.39046 47.856056 50.29154 22.796968 49.55857
52.2755 49.18443 54.541786 49.078922 52.96913 47.357506 49.91127
40.52969 48.609577 46.219105 48.567707 47.50982 50.94427 49.646065
50.107346 50.30451 47.274677 51.30897 51.84024 ]
Category counts: [ 11 13 1 21 211 14 17 30 18 630 18 15 1250 11
7 169 16 10 28 1375 0 16 12 163 1 54 19 13
692 1 61 1 15 5 6 10 7 14 20 25]
New alpha: [1.1312500e+01 1.3312500e+01 1.3125000e+00 2.1312500e+01 2.1131250e+02
1.4312500e+01 1.7312500e+01 3.0312500e+01 1.8312500e+01 6.3031250e+02
1.8312500e+01 1.5312500e+01 1.2503125e+03 1.1312500e+01 7.3125000e+00
1.6931250e+02 1.6312500e+01 1.0312500e+01 2.8312500e+01 1.3753125e+03
3.1250000e-01 1.6312500e+01 1.2312500e+01 1.6331250e+02 1.3125000e+00
5.4312500e+01 1.9312500e+01 1.3312500e+01 6.9231250e+02 1.3125000e+00
6.1312500e+01 1.3125000e+00 1.5312500e+01 5.3125000e+00 6.3125000e+00
1.0312500e+01 7.3125000e+00 1.4312500e+01 2.0312500e+01 2.5312500e+01]
New probs: [2.1673818e-03 1.7163822e-03 8.1811310e-04 4.2350576e-03 4.2799558e-02
2.7764970e-03 3.9303540e-03 5.7810517e-03 2.9422438e-03 1.2634218e-01
2.8213260e-03 2.7249898e-03 2.5118956e-01 2.5427374e-03 1.6928094e-03
3.7217326e-02 4.6840007e-03 2.8148382e-03 4.7610039e-03 2.7655897e-01
2.5362361e-04 3.4750560e-03 2.9044515e-03 2.7890552e-02 2.7338593e-04
1.1126963e-02 3.3351232e-03 2.2516837e-03 1.3335088e-01 1.6660364e-04
1.3140402e-02 1.4395178e-04 3.4992164e-03 1.5792440e-03 2.1839882e-03
2.2756658e-03 1.1726766e-03 2.0768612e-03 4.6465592e-03 3.7363614e-03]
Category counts: [ 11 13 1 21 211 14 17 30 18 630 18 15 1250 11
7 169 16 10 28 1375 0 16 12 163 1 54 19 13
692 1 61 1 15 5 6 10 7 14 20 25]
Current means: [47.39785 51.875507 50.21505 47.12551 57.647923 51.44297 50.916122
48.02084 48.269955 64.443245 48.291588 50.655987 79.48817 51.132618
48.342808 43.86549 47.91075 47.37451 50.184063 20.496754 nan
52.266758 49.01793 54.63084 48.333443 52.93877 46.93995 49.607124
37.093113 46.40639 46.192722 48.725506 47.59996 50.771637 49.858982
50.18091 50.3991 46.710514 51.819206 52.083218]
Posterior variance: [9.0090089e-02 7.6335877e-02 9.0909088e-01 4.7393363e-02 4.7370913e-03
7.0921987e-02 5.8479533e-02 3.3222590e-02 5.5248618e-02 1.5870497e-03
5.5248618e-02 6.6225164e-02 7.9993601e-04 9.0090089e-02 1.4084508e-01
5.9136604e-03 6.2111799e-02 9.9009894e-02 3.5587188e-02 7.2721986e-04
1.0000000e+01 6.2111799e-02 8.2644626e-02 6.1312076e-03 9.0909088e-01
1.8484289e-02 5.2356020e-02 7.6335877e-02 1.4448779e-03 9.0909088e-01
1.6366612e-02 9.0909088e-01 6.6225164e-02 1.9607843e-01 1.6393442e-01
9.9009894e-02 1.4084508e-01 7.0921987e-02 4.9751244e-02 3.9840639e-02]
Posterior means: [47.421295 51.861187 50.195503 47.139133 57.6443 51.43274 50.910767
48.027416 48.279514 64.440956 48.301025 50.651646 79.48581 51.122414
48.366146 43.869114 47.92373 47.4005 50.18341 20.498898 nan
52.252678 49.026047 54.628002 48.48495 52.933342 46.95597 49.610123
37.09498 46.733086 46.19895 48.841373 47.615856 50.756508 49.86129
50.17911 50.39348 46.733845 51.810154 52.07492 ]
Resampled means: [47.30411 52.00231 50.446846 47.061993 57.644936 51.527733 50.987686
48.058838 48.407818 64.44234 48.312874 50.54396 79.48622 51.04276
48.500233 43.873783 47.821304 47.39117 50.199932 20.499535 nan
52.259933 48.932796 54.621555 49.038933 52.931484 46.953037 49.53742
37.092857 45.87163 46.179974 49.528522 47.7202 50.746693 50.02338
50.251102 50.572582 46.671837 51.988567 52.059753]
Chosen means: [47.30411 52.00231 50.446846 47.061993 57.644936 51.527733 50.987686
48.058838 48.407818 64.44234 48.312874 50.54396 79.48622 51.04276
48.500233 43.873783 47.821304 47.39117 50.199932 20.499535 49.55857
52.259933 48.932796 54.621555 49.038933 52.931484 46.953037 49.53742
37.092857 45.87163 46.179974 49.528522 47.7202 50.746693 50.02338
50.251102 50.572582 46.671837 51.988567 52.059753]
Category counts: [ 7 9 1 28 323 17 28 31 14 659 16 18 1118 17
9 289 18 12 18 1230 0 18 14 154 1 59 15 6
720 1 66 1 12 9 11 9 4 8 20 10]
New alpha: [7.3125000e+00 9.3125000e+00 1.3125000e+00 2.8312500e+01 3.2331250e+02
1.7312500e+01 2.8312500e+01 3.1312500e+01 1.4312500e+01 6.5931250e+02
1.6312500e+01 1.8312500e+01 1.1183125e+03 1.7312500e+01 9.3125000e+00
2.8931250e+02 1.8312500e+01 1.2312500e+01 1.8312500e+01 1.2303125e+03
3.1250000e-01 1.8312500e+01 1.4312500e+01 1.5431250e+02 1.3125000e+00
5.9312500e+01 1.5312500e+01 6.3125000e+00 7.2031250e+02 1.3125000e+00
6.6312500e+01 1.3125000e+00 1.2312500e+01 9.3125000e+00 1.1312500e+01
9.3125000e+00 4.3125000e+00 8.3125000e+00 2.0312500e+01 1.0312500e+01]
New probs: [1.0484123e-03 2.4039901e-03 3.2780448e-04 8.9008873e-03 6.3795947e-02
4.2834594e-03 5.2645109e-03 8.5770758e-03 2.4766033e-03 1.3452582e-01
2.4886839e-03 4.1137994e-03 2.2325648e-01 4.6034381e-03 2.2076806e-03
5.7211764e-02 3.2629785e-03 2.6335120e-03 3.9636041e-03 2.4250145e-01
2.1505499e-05 2.9059877e-03 1.7649684e-03 2.7178811e-02 4.8947573e-04
1.3483511e-02 3.8566892e-03 1.2811716e-03 1.4238575e-01 4.0745517e-04
1.2738028e-02 2.7574910e-04 2.9633243e-03 1.9458992e-03 1.9879607e-03
1.0513082e-03 4.9322203e-04 1.1878294e-03 4.2131674e-03 1.5200591e-03]
Category counts: [ 7 9 1 28 323 17 28 31 14 659 16 18 1118 17
9 289 18 12 18 1230 0 18 14 154 1 59 15 6
720 1 66 1 12 9 11 9 4 8 20 10]
Current means: [47.5665 51.656178 50.594707 46.927044 58.47989 51.27839 50.795383
47.922024 48.672974 66.31128 48.6865 50.697826 80.53894 51.196953
48.528244 43.01227 47.92112 47.435852 50.304123 19.346678 nan
52.62142 49.017117 54.580574 50.893482 52.66872 46.8165 50.146633
34.965893 42.873196 46.31348 50.494926 47.79755 50.269093 49.98281
49.804783 49.90245 46.301067 51.882484 51.96656 ]
Posterior variance: [1.4084508e-01 1.0989010e-01 9.0909088e-01 3.5587188e-02 3.0950170e-03
5.8479533e-02 3.5587188e-02 3.2154340e-02 7.0921987e-02 1.5172205e-03
6.2111799e-02 5.5248618e-02 8.9437439e-04 5.8479533e-02 1.0989010e-01
3.4590107e-03 5.5248618e-02 8.2644626e-02 5.5248618e-02 8.1294205e-04
1.0000000e+01 5.5248618e-02 7.0921987e-02 6.4892923e-03 9.0909088e-01
1.6920473e-02 6.6225164e-02 1.6393442e-01 1.3886960e-03 9.0909088e-01
1.5128593e-02 9.0909088e-01 8.2644626e-02 1.0989010e-01 9.0090089e-02
1.0989010e-01 2.4390244e-01 1.2345678e-01 4.9751244e-02 9.9009894e-02]
Posterior means: [47.600773 51.63798 50.540646 46.93798 58.47727 51.270912 50.792553
47.928703 48.68239 66.30881 48.69466 50.69397 80.53621 51.189953
48.54442 43.01469 47.932606 47.457047 50.302444 19.349169 nan
52.60694 49.02409 54.577602 50.81226 52.664207 46.837585 50.144226
34.967983 43.521088 46.319054 50.449936 47.815754 50.266136 49.982967
49.80693 49.90483 46.346733 51.87312 51.947086]
Resampled means: [47.828007 51.598907 50.26712 46.960922 58.4742 51.2921 50.80302
47.98146 48.73236 66.30938 48.726753 50.662422 80.5354 51.12578
48.52776 43.012943 47.939507 47.497765 50.337616 19.347212 nan
52.545887 48.94067 54.578136 51.107822 52.66197 46.845142 50.30564
34.97025 45.91524 46.317657 50.1744 47.79839 50.145996 49.88021
49.73512 50.241524 46.441093 51.926933 52.063957]
Chosen means: [47.828007 51.598907 50.26712 46.960922 58.4742 51.2921 50.80302
47.98146 48.73236 66.30938 48.726753 50.662422 80.5354 51.12578
48.52776 43.012943 47.939507 47.497765 50.337616 19.347212 49.55857
52.545887 48.94067 54.578136 51.107822 52.66197 46.845142 50.30564
34.97025 45.91524 46.317657 50.1744 47.79839 50.145996 49.88021
49.73512 50.241524 46.441093 51.926933 52.063957]
Category counts: [ 6 6 0 47 373 18 26 36 17 709 11 24 1000 16
9 360 14 13 17 1124 0 19 13 168 3 72 16 5
745 4 63 3 10 10 5 5 5 5 20 3]
New alpha: [6.3125000e+00 6.3125000e+00 3.1250000e-01 4.7312500e+01 3.7331250e+02
1.8312500e+01 2.6312500e+01 3.6312500e+01 1.7312500e+01 7.0931250e+02
1.1312500e+01 2.4312500e+01 1.0003125e+03 1.6312500e+01 9.3125000e+00
3.6031250e+02 1.4312500e+01 1.3312500e+01 1.7312500e+01 1.1243125e+03
3.1250000e-01 1.9312500e+01 1.3312500e+01 1.6831250e+02 3.3125000e+00
7.2312500e+01 1.6312500e+01 5.3125000e+00 7.4531250e+02 4.3125000e+00
6.3312500e+01 3.3125000e+00 1.0312500e+01 1.0312500e+01 5.3125000e+00
5.3125000e+00 5.3125000e+00 5.3125000e+00 2.0312500e+01 3.3125000e+00]
New probs: [7.1326061e-04 1.0461451e-03 1.0887781e-08 8.2955621e-03 7.9418242e-02
3.9928402e-03 4.7319275e-03 6.1478820e-03 3.6614372e-03 1.3740961e-01
2.5528162e-03 3.5433522e-03 1.9913024e-01 3.7447643e-03 1.4898101e-03
7.3318720e-02 3.2863538e-03 2.6568412e-03 2.8536820e-03 2.2291099e-01
1.0118150e-05 4.0252367e-03 4.0440694e-03 3.3246946e-02 3.4518112e-04
1.8657612e-02 3.3503885e-03 1.0110930e-03 1.4722630e-01 1.4468585e-03
1.0326716e-02 7.4384233e-04 1.2104714e-03 2.9335415e-03 9.4662694e-04
1.3073983e-03 1.9155390e-03 2.3673971e-03 3.3520185e-03 6.2791962e-04]
Category counts: [ 6 6 0 47 373 18 26 36 17 709 11 24 1000 16
9 360 14 13 17 1124 0 19 13 168 3 72 16 5
745 4 63 3 10 10 5 5 5 5 20 3]
Current means: [47.332397 51.630253 nan 46.729717 59.172085 51.16313 50.84512
48.33057 48.83966 67.79423 48.690014 50.25913 81.48343 51.147327
48.43475 42.31971 47.59215 47.22457 50.255646 18.492432 nan
52.605515 49.396225 54.845768 51.324566 52.432114 46.922245 50.947887
33.459164 45.659206 46.20593 50.318787 47.743675 49.9695 49.078297
49.833294 50.4654 46.65812 52.149376 52.274017]
Posterior variance: [1.63934425e-01 1.63934425e-01 1.00000000e+01 2.12314241e-02
2.68024649e-03 5.52486181e-02 3.83141749e-02 2.77008321e-02
5.84795326e-02 1.41023833e-03 9.00900885e-02 4.14937735e-02
9.99900047e-04 6.21117987e-02 1.09890103e-01 2.77700624e-03
7.09219873e-02 7.63358772e-02 5.84795326e-02 8.89600604e-04
1.00000000e+01 5.23560196e-02 7.63358772e-02 5.94883971e-03
3.22580665e-01 1.38696255e-02 6.21117987e-02 1.96078435e-01
1.34210172e-03 2.43902445e-01 1.58478618e-02 3.22580665e-01
9.90098938e-02 9.90098938e-02 1.96078435e-01 1.96078435e-01
1.96078435e-01 1.96078435e-01 4.97512445e-02 3.22580665e-01]
Posterior means: [47.376125 51.603523 nan 46.73666 59.16963 51.156708 50.84188
48.335194 48.846447 67.79172 48.701817 50.258053 81.48028 51.1402
48.451954 42.321842 47.60923 47.245754 50.25415 18.495234 nan
52.591877 49.400833 54.842884 51.28184 52.428738 46.94136 50.929302
33.461384 45.76508 46.21194 50.308506 47.76601 49.9698 49.09637
49.836563 50.456276 46.723648 52.138683 52.200665]
Resampled means: [47.49356 51.927406 nan 46.71819 59.171284 51.177567 50.878338
48.351284 48.744583 67.79259 48.54787 50.362663 81.47964 51.191933
48.401867 42.3215 47.586163 47.22909 50.38704 18.494173 nan
52.625637 49.293224 54.84654 51.34701 52.435898 46.98339 50.868553
33.461903 46.008972 46.20115 49.932728 47.75515 49.90669 48.81847
49.526375 50.49478 46.769543 52.16988 52.354877]
Chosen means: [47.49356 51.927406 50.26712 46.71819 59.171284 51.177567 50.878338
48.351284 48.744583 67.79259 48.54787 50.362663 81.47964 51.191933
48.401867 42.3215 47.586163 47.22909 50.38704 18.494173 49.55857
52.625637 49.293224 54.84654 51.34701 52.435898 46.98339 50.868553
33.461903 46.008972 46.20115 49.932728 47.75515 49.90669 48.81847
49.526375 50.49478 46.769543 52.16988 52.354877]
Category counts: [ 1 3 0 44 407 21 21 34 22 704 16 18 921 18
6 448 19 8 18 1004 0 11 22 216 1 90 24 4
753 9 63 7 11 12 3 3 7 13 15 3]
New alpha: [1.3125000e+00 3.3125000e+00 3.1250000e-01 4.4312500e+01 4.0731250e+02
2.1312500e+01 2.1312500e+01 3.4312500e+01 2.2312500e+01 7.0431250e+02
1.6312500e+01 1.8312500e+01 9.2131250e+02 1.8312500e+01 6.3125000e+00
4.4831250e+02 1.9312500e+01 8.3125000e+00 1.8312500e+01 1.0043125e+03
3.1250000e-01 1.1312500e+01 2.2312500e+01 2.1631250e+02 1.3125000e+00
9.0312500e+01 2.4312500e+01 4.3125000e+00 7.5331250e+02 9.3125000e+00
6.3312500e+01 7.3125000e+00 1.1312500e+01 1.2312500e+01 3.3125000e+00
3.3125000e+00 7.3125000e+00 1.3312500e+01 1.5312500e+01 3.3125000e+00]
New probs: [2.1196724e-04 4.5253025e-04 3.2575281e-05 9.5339343e-03 8.0842450e-02
3.0001272e-03 4.3563158e-03 5.6997277e-03 4.6140766e-03 1.3624157e-01
2.9122075e-03 5.5064419e-03 1.7984831e-01 2.4481947e-03 1.2753750e-03
9.3407750e-02 4.9266312e-03 1.0861974e-03 4.7688335e-03 2.0985861e-01
4.8795211e-04 2.4426975e-03 4.3230578e-03 3.9631508e-02 7.3840463e-04
2.0077281e-02 4.4969236e-03 4.2944416e-04 1.4620233e-01 2.2138201e-03
1.1991677e-02 1.5104315e-03 1.9393002e-03 2.5195284e-03 5.1665300e-04
8.5688470e-04 1.5809762e-03 2.1947860e-03 3.8904811e-03 9.3227089e-04]
Category counts: [ 1 3 0 44 407 21 21 34 22 704 16 18 921 18
6 448 19 8 18 1004 0 11 22 216 1 90 24 4
753 9 63 7 11 12 3 3 7 13 15 3]
Current means: [46.974678 50.026287 nan 46.330044 60.208897 51.06257 50.896297
48.457096 48.689053 69.1429 47.95723 50.31358 82.09974 51.147327
48.495136 41.245296 47.42284 46.71911 50.480453 17.532934 nan
52.364307 49.29354 55.23518 52.573154 52.56875 46.67065 51.397575
31.597698 45.965088 45.92332 49.49292 47.225513 50.091312 48.422256
49.780476 50.550938 46.400322 51.988346 51.72092 ]
Posterior variance: [9.0909088e-01 3.2258067e-01 1.0000000e+01 2.2675738e-02 2.4563989e-03
4.7393363e-02 4.7393363e-02 2.9325515e-02 4.5248870e-02 1.4202528e-03
6.2111799e-02 5.5248618e-02 1.0856584e-03 5.5248618e-02 1.6393442e-01
2.2316447e-03 5.2356020e-02 1.2345678e-01 5.5248618e-02 9.9591678e-04
1.0000000e+01 9.0090089e-02 4.5248870e-02 4.6274872e-03 9.0909088e-01
1.1098780e-02 4.1493773e-02 2.4390244e-01 1.3278449e-03 1.0989010e-01
1.5847862e-02 1.4084508e-01 9.0090089e-02 8.2644626e-02 3.2258067e-01
3.2258067e-01 1.4084508e-01 7.6335877e-02 6.6225164e-02 3.2258067e-01]
Posterior means: [47.24971 50.02544 nan 46.338367 60.20639 51.05753 50.892048
48.461624 48.694984 69.14018 47.969917 50.311848 82.096245 51.140987
48.519802 41.24725 47.436337 46.759613 50.4778 17.53617 nan
52.34301 49.296738 55.232754 52.339233 52.565895 46.684464 51.36349
31.60014 46.00943 45.929783 49.50006 47.25051 50.09056 48.473152
49.78756 50.543175 46.4278 51.97518 51.66541 ]
Resampled means: [46.191177 49.97323 nan 46.353462 60.205658 51.04557 50.897343
48.459435 48.68233 69.139824 47.973633 50.211475 82.0926 51.193787
48.434708 41.24564 47.455452 46.851673 50.439686 17.537254 nan
52.552547 49.30614 55.232113 51.906845 52.566788 46.657997 51.151287
31.599327 46.119785 45.91741 49.408722 47.387657 50.13795 48.686172
49.945545 50.40278 46.420628 52.03433 51.47237 ]
Chosen means: [46.191177 49.97323 50.26712 46.353462 60.205658 51.04557 50.897343
48.459435 48.68233 69.139824 47.973633 50.211475 82.0926 51.193787
48.434708 41.24564 47.455452 46.851673 50.439686 17.537254 49.55857
52.552547 49.30614 55.232113 51.906845 52.566788 46.657997 51.151287
31.599327 46.119785 45.91741 49.408722 47.387657 50.13795 48.686172
49.945545 50.40278 46.420628 52.03433 51.47237 ]
Category counts: [ 1 4 0 58 455 17 27 37 30 669 19 24 875 9 4 451 27 3
23 952 2 21 24 233 2 95 22 1 730 14 106 6 13 7 3 5
4 6 13 8]
New alpha: [1.312500e+00 4.312500e+00 3.125000e-01 5.831250e+01 4.553125e+02
1.731250e+01 2.731250e+01 3.731250e+01 3.031250e+01 6.693125e+02
1.931250e+01 2.431250e+01 8.753125e+02 9.312500e+00 4.312500e+00
4.513125e+02 2.731250e+01 3.312500e+00 2.331250e+01 9.523125e+02
2.312500e+00 2.131250e+01 2.431250e+01 2.333125e+02 2.312500e+00
9.531250e+01 2.231250e+01 1.312500e+00 7.303125e+02 1.431250e+01
1.063125e+02 6.312500e+00 1.331250e+01 7.312500e+00 3.312500e+00
5.312500e+00 4.312500e+00 6.312500e+00 1.331250e+01 8.312500e+00]
New probs: [1.12070746e-04 5.61572437e-04 1.39575477e-07 1.21237952e-02
8.70418549e-02 4.88283578e-03 5.92409726e-03 8.17592442e-03
6.22691307e-03 1.32962987e-01 5.61798224e-03 4.72867256e-03
1.65950254e-01 7.79046968e-04 1.38890930e-03 9.07041878e-02
5.62213594e-03 8.65971844e-04 4.28505894e-03 1.88465029e-01
2.35402054e-04 4.50017350e-03 3.95431463e-03 4.88412604e-02
5.88222058e-04 2.07980461e-02 2.95081455e-03 1.99118542e-04
1.52881265e-01 2.08916282e-03 2.00091079e-02 1.72951492e-03
2.81626149e-03 1.01801055e-03 1.47231622e-04 9.92273679e-04
4.46267310e-04 2.30870559e-03 4.69900295e-03 2.37620948e-03]
Category counts: [ 1 4 0 58 455 17 27 37 30 669 19 24 875 9 4 451 27 3
23 952 2 21 24 233 2 95 22 1 730 14 106 6 13 7 3 5
4 6 13 8]
Current means: [46.553684 50.509796 nan 45.647263 61.186718 50.98982 50.83519
48.41371 48.720894 70.11232 47.65002 50.29746 82.48516 51.558365
48.15818 40.010532 47.10794 46.20149 50.904076 17.134108 49.779804
52.887093 49.145176 55.557438 53.173378 52.63635 46.29965 50.95144
30.572725 45.33523 45.188854 49.480473 47.320698 49.749287 48.55322
50.062363 49.96749 46.715595 51.81576 51.938717]
Posterior variance: [9.09090877e-01 2.43902445e-01 1.00000000e+01 1.72117036e-02
2.19731918e-03 5.84795326e-02 3.69003676e-02 2.69541796e-02
3.32225896e-02 1.49454491e-03 5.23560196e-02 4.14937735e-02
1.14272663e-03 1.09890103e-01 2.43902445e-01 2.21680338e-03
3.69003676e-02 3.22580665e-01 4.32900414e-02 1.05030986e-03
4.76190507e-01 4.73933630e-02 4.14937735e-02 4.29000426e-03
4.76190507e-01 1.05152475e-02 4.52488698e-02 9.09090877e-01
1.36967539e-03 7.09219873e-02 9.42507107e-03 1.63934425e-01
7.63358772e-02 1.40845075e-01 3.22580665e-01 1.96078435e-01
2.43902445e-01 1.63934425e-01 7.63358772e-02 1.23456784e-01]
Posterior means: [46.86699 50.497364 nan 45.654755 61.184258 50.98403 50.832108
48.417984 48.725143 70.109314 47.662327 50.296223 82.48145 51.54124
48.203106 40.01275 47.118614 46.324024 50.90016 17.13756 49.790287
52.87341 49.14872 55.555054 53.022263 52.633575 46.31639 50.86495
30.575388 45.368317 45.193394 49.488987 47.34115 49.752815 48.59989
50.061142 49.968285 46.769436 51.8019 51.91478 ]
Resampled means: [47.24992 50.43668 nan 45.64226 61.187237 50.857918 50.85741
48.43646 48.787346 70.10977 47.736137 50.21902 82.480835 51.725555
48.362053 40.01132 47.12486 46.671288 50.968216 17.138233 48.73583
52.894943 49.14806 55.55691 53.2842 52.63431 46.26711 51.45684
30.575949 45.299133 45.181717 49.6406 47.395813 49.945343 48.853664
49.918262 50.107838 46.749577 51.855827 51.95197 ]
Chosen means: [47.24992 50.43668 50.26712 45.64226 61.187237 50.857918 50.85741
48.43646 48.787346 70.10977 47.736137 50.21902 82.480835 51.725555
48.362053 40.01132 47.12486 46.671288 50.968216 17.138233 48.73583
52.894943 49.14806 55.55691 53.2842 52.63431 46.26711 51.45684
30.575949 45.299133 45.181717 49.6406 47.395813 49.945343 48.853664
49.918262 50.107838 46.749577 51.855827 51.95197 ]
Category counts: [ 1 7 0 58 466 21 25 38 26 662 28 23 838 4 3 472 24 4
18 879 3 20 20 264 1 95 14 1 746 16 132 12 17 9 1 9
2 8 18 15]
New alpha: [1.312500e+00 7.312500e+00 3.125000e-01 5.831250e+01 4.663125e+02
2.131250e+01 2.531250e+01 3.831250e+01 2.631250e+01 6.623125e+02
2.831250e+01 2.331250e+01 8.383125e+02 4.312500e+00 3.312500e+00
4.723125e+02 2.431250e+01 4.312500e+00 1.831250e+01 8.793125e+02
3.312500e+00 2.031250e+01 2.031250e+01 2.643125e+02 1.312500e+00
9.531250e+01 1.431250e+01 1.312500e+00 7.463125e+02 1.631250e+01
1.323125e+02 1.231250e+01 1.731250e+01 9.312500e+00 1.312500e+00
9.312500e+00 2.312500e+00 8.312500e+00 1.831250e+01 1.531250e+01]
New probs: [1.7255213e-04 1.7437905e-03 7.0047790e-05 1.3498584e-02 8.8707238e-02
4.2700893e-03 3.6912155e-03 9.9082692e-03 5.0118752e-03 1.2570179e-01
4.0599168e-03 4.5133037e-03 1.7171872e-01 1.4297430e-03 4.0831853e-04
9.4778776e-02 4.4669481e-03 6.7734107e-04 4.7777286e-03 1.8098311e-01
8.7095948e-04 2.7769036e-03 4.8450013e-03 5.0776511e-02 7.5217139e-04
1.8370463e-02 1.5096143e-03 2.7587797e-04 1.5220124e-01 4.0762364e-03
2.5588537e-02 1.8118835e-03 3.4500062e-03 1.6716123e-03 4.7863432e-05
1.5112355e-03 1.2982346e-04 1.3468731e-03 4.7776964e-03 2.6198598e-03]
Category counts: [ 1 7 0 58 466 21 25 38 26 662 28 23 838 4 3 472 24 4
18 879 3 20 20 264 1 95 14 1 746 16 132 12 17 9 1 9
2 8 18 15]
Current means: [46.96401 50.630222 nan 45.33648 61.732082 50.724968 50.865448
48.691673 48.56413 70.81073 47.71924 50.213295 82.76319 51.999947
47.366413 39.325752 46.892506 47.093456 50.669594 16.53939 48.49992
53.00602 48.533047 55.919613 54.705997 52.490086 45.693462 50.984196
29.522243 45.14081 44.61482 49.679348 47.518337 49.202652 50.338783
50.292072 49.750435 46.436775 51.89885 52.41743 ]
Posterior variance: [9.09090877e-01 1.40845075e-01 1.00000000e+01 1.72117036e-02
2.14546220e-03 4.73933630e-02 3.98406386e-02 2.62467209e-02
3.83141749e-02 1.51034596e-03 3.55871879e-02 4.32900414e-02
1.19317509e-03 2.43902445e-01 3.22580665e-01 2.11819517e-03
4.14937735e-02 2.43902445e-01 5.52486181e-02 1.13752706e-03
3.22580665e-01 4.97512445e-02 4.97512445e-02 3.78644443e-03
9.09090877e-01 1.05152475e-02 7.09219873e-02 9.09090877e-01
1.34030299e-03 6.21117987e-02 7.57002225e-03 8.26446265e-02
5.84795326e-02 1.09890103e-01 9.09090877e-01 1.09890103e-01
4.76190507e-01 1.23456784e-01 5.52486181e-02 6.62251636e-02]
Posterior means: [47.24001 50.621346 nan 45.344505 61.72957 50.72153 50.862003
48.695107 48.56963 70.80759 47.727356 50.21237 82.759285 51.951168
47.45137 39.32801 46.9054 47.16435 50.665894 16.543196 48.548313
52.991062 48.540344 55.91737 54.278183 52.48747 45.724007 50.894726
29.524988 45.17099 44.618893 49.682 47.53285 49.21142 50.307987
50.288864 49.762318 46.480766 51.88836 52.401424]
Resampled means: [47.543415 50.35954 nan 45.337807 61.727367 50.75066 50.82397
48.68751 48.58279 70.809944 47.73279 50.184345 82.75879 52.207493
47.32297 39.327106 46.952652 47.227066 50.734917 16.544348 48.26844
52.821487 48.57457 55.915943 55.76784 52.493423 45.68177 51.161404
29.524757 45.180626 44.61609 49.798714 47.480343 49.214035 50.325737
50.26275 50.75731 46.38484 52.03652 52.479916]
Chosen means: [47.543415 50.35954 50.26712 45.337807 61.727367 50.75066 50.82397
48.68751 48.58279 70.809944 47.73279 50.184345 82.75879 52.207493
47.32297 39.327106 46.952652 47.227066 50.734917 16.544348 48.26844
52.821487 48.57457 55.915943 55.76784 52.493423 45.68177 51.161404
29.524757 45.180626 44.61609 49.798714 47.480343 49.214035 50.325737
50.26275 50.75731 46.38484 52.03652 52.479916]
Category counts: [ 1 8 2 73 448 19 18 47 23 691 25 24 779 6 5 480 28 3
31 870 6 17 15 310 4 86 11 2 685 19 175 10 18 8 0 6
1 7 25 14]
New alpha: [1.312500e+00 8.312500e+00 2.312500e+00 7.331250e+01 4.483125e+02
1.931250e+01 1.831250e+01 4.731250e+01 2.331250e+01 6.913125e+02
2.531250e+01 2.431250e+01 7.793125e+02 6.312500e+00 5.312500e+00
4.803125e+02 2.831250e+01 3.312500e+00 3.131250e+01 8.703125e+02
6.312500e+00 1.731250e+01 1.531250e+01 3.103125e+02 4.312500e+00
8.631250e+01 1.131250e+01 2.312500e+00 6.853125e+02 1.931250e+01
1.753125e+02 1.031250e+01 1.831250e+01 8.312500e+00 3.125000e-01
6.312500e+00 1.312500e+00 7.312500e+00 2.531250e+01 1.431250e+01]
New probs: [9.75410658e-05 1.90390670e-03 2.84772133e-04 1.50025031e-02
9.22242180e-02 2.42311996e-03 3.99559783e-03 9.00209788e-03
5.70507348e-03 1.34769231e-01 6.61733467e-03 4.76223929e-03
1.55421838e-01 9.98117728e-04 6.55599288e-04 9.41333994e-02
5.23053575e-03 3.28179216e-04 8.04566126e-03 1.77873328e-01
1.89502956e-03 3.00383172e-03 3.42095038e-03 6.20650053e-02
4.35220252e-04 1.80077255e-02 3.31702223e-03 1.14749106e-04
1.29299209e-01 4.52292338e-03 3.37460563e-02 1.87533966e-03
4.67356667e-03 1.38486072e-03 1.67539947e-05 1.77470304e-03
1.19174285e-04 1.53805560e-03 4.72347159e-03 4.59202426e-03]
Category counts: [ 1 8 2 73 448 19 18 47 23 691 25 24 779 6 5 480 28 3
31 870 6 17 15 310 4 86 11 2 685 19 175 10 18 8 0 6
1 7 25 14]
Current means: [48.44412 50.327583 49.621346 45.262417 62.378178 50.256073 50.71908
48.504234 48.542152 71.49419 47.50138 50.49149 83.24108 52.364574
48.563988 38.242016 47.051888 47.920345 50.97668 16.468143 47.708622
52.87016 48.46423 56.310257 54.86656 52.66448 45.939266 51.525986
28.911205 44.72741 43.821114 49.908104 47.459503 49.24861 nan
50.491776 52.178345 46.398464 51.676743 52.478638]
Posterior variance: [9.0909088e-01 1.2345678e-01 4.7619051e-01 1.3679891e-02 2.2316447e-03
5.2356020e-02 5.5248618e-02 2.1231424e-02 4.3290041e-02 1.4469686e-03
3.9840639e-02 4.1493773e-02 1.2835324e-03 1.6393442e-01 1.9607843e-01
2.0828994e-03 3.5587188e-02 3.2258067e-01 3.2154340e-02 1.1492933e-03
1.6393442e-01 5.8479533e-02 6.6225164e-02 3.2247661e-03 2.4390244e-01
1.1614402e-02 9.0090089e-02 4.7619051e-01 1.4596410e-03 5.2356020e-02
5.7110223e-03 9.9009894e-02 5.5248618e-02 1.2345678e-01 1.0000000e+01
1.6393442e-01 9.0909088e-01 1.4084508e-01 3.9840639e-02 7.0921987e-02]
Posterior means: [48.585564 50.32354 49.639374 45.268898 62.375412 50.254734 50.71511
48.50741 48.548466 71.49107 47.511337 50.48945 83.23682 52.32581
48.592144 38.244465 47.06238 47.987434 50.973537 16.471998 47.74618
52.853374 48.474403 56.30822 54.747868 52.66138 45.97585 51.45332
28.914282 44.755016 43.824642 49.90901 47.47354 49.25789 nan
50.48371 51.980316 46.44919 51.670063 52.46106 ]
Resampled means: [49.972168 50.368374 49.871727 45.270363 62.374104 50.256958 50.630146
48.491043 48.519253 71.492134 47.52078 50.516293 83.23463 52.375313
48.763275 38.246532 47.056793 48.232246 50.977425 16.472376 47.762177
52.806744 48.445076 56.307068 54.842422 52.6551 45.85516 50.80916
28.91604 44.659687 43.820034 49.84225 47.51146 49.24012 nan
50.337452 51.840244 46.436718 51.645786 52.50118 ]
Chosen means: [49.972168 50.368374 49.871727 45.270363 62.374104 50.256958 50.630146
48.491043 48.519253 71.492134 47.52078 50.516293 83.23463 52.375313
48.763275 38.246532 47.056793 48.232246 50.977425 16.472376 47.762177
52.806744 48.445076 56.307068 54.842422 52.6551 45.85516 50.80916
28.91604 44.659687 43.820034 49.84225 47.51146 49.24012 50.325737
50.337452 51.840244 46.436718 51.645786 52.50118 ]
Category counts: [ 1 8 1 64 485 10 20 42 35 640 27 24 750 2 1 494 30 2
39 840 6 10 8 335 2 116 13 0 660 19 208 11 29 7 1 11
0 9 13 27]
New alpha: [1.312500e+00 8.312500e+00 1.312500e+00 6.431250e+01 4.853125e+02
1.031250e+01 2.031250e+01 4.231250e+01 3.531250e+01 6.403125e+02
2.731250e+01 2.431250e+01 7.503125e+02 2.312500e+00 1.312500e+00
4.943125e+02 3.031250e+01 2.312500e+00 3.931250e+01 8.403125e+02
6.312500e+00 1.031250e+01 8.312500e+00 3.353125e+02 2.312500e+00
1.163125e+02 1.331250e+01 3.125000e-01 6.603125e+02 1.931250e+01
2.083125e+02 1.131250e+01 2.931250e+01 7.312500e+00 1.312500e+00
1.131250e+01 3.125000e-01 9.312500e+00 1.331250e+01 2.731250e+01]
New probs: [5.7480967e-04 1.4994700e-03 3.0095348e-04 1.1686050e-02 9.5154449e-02
1.9017545e-03 3.4956592e-03 1.0954513e-02 6.5235398e-03 1.3139531e-01
5.4558325e-03 6.0229660e-03 1.5705052e-01 2.5057344e-04 1.4572061e-04
9.3264803e-02 5.3448556e-03 5.4595462e-04 7.4963598e-03 1.5684442e-01
1.1840909e-03 2.3546708e-03 1.2630308e-03 7.3216803e-02 6.0553622e-04
2.3319080e-02 3.0539313e-03 7.5271746e-05 1.2869376e-01 5.3290543e-03
4.0202804e-02 2.4825064e-03 7.3668319e-03 4.5061536e-04 6.8469199e-05
2.3138809e-03 1.5969066e-05 2.6252246e-03 2.5482019e-03 6.9222078e-03]
Category counts: [ 1 8 1 64 485 10 20 42 35 640 27 24 750 2 1 494 30 2
39 840 6 10 8 335 2 116 13 0 660 19 208 11 29 7 1 11
0 9 13 27]
Current means: [50.66751 50.205837 50.733574 45.3675 63.381996 50.36564 50.38019
48.3326 48.434235 72.35693 47.442207 50.506763 83.48977 52.467308
48.918564 37.487213 46.88409 48.437355 50.872704 16.240767 48.08356
52.83175 48.67267 56.745087 54.70683 52.753716 46.625122 nan
28.186459 44.79764 43.25035 50.05466 47.195625 48.468727 50.49217
50.87903 nan 46.58785 51.59312 52.965263]
Posterior variance: [9.09090877e-01 1.23456784e-01 9.09090877e-01 1.56006245e-02
2.06143060e-03 9.90098938e-02 4.97512445e-02 2.37529706e-02
2.84900293e-02 1.56225590e-03 3.69003676e-02 4.14937735e-02
1.33315567e-03 4.76190507e-01 9.09090877e-01 2.02388177e-03
3.32225896e-02 4.76190507e-01 2.55754478e-02 1.19033456e-03
1.63934425e-01 9.90098938e-02 1.23456784e-01 2.98418384e-03
4.76190507e-01 8.61326419e-03 7.63358772e-02 1.00000000e+01
1.51492201e-03 5.23560196e-02 4.80538188e-03 9.00900885e-02
3.43642607e-02 1.40845075e-01 9.09090877e-01 9.00900885e-02
1.00000000e+01 1.09890103e-01 7.63358772e-02 3.69003676e-02]
Posterior means: [50.60683 50.203297 50.66689 45.374725 63.37924 50.362015 50.3783
48.33656 48.438694 72.35345 47.45165 50.504658 83.485306 52.349815
49.01688 37.489746 46.894444 48.511765 50.87047 16.244783 48.114975
52.803707 48.689056 56.743073 54.482693 52.751347 46.650883 nan
28.189762 44.824883 43.253597 50.05417 47.20526 48.49029 50.44743
50.87111 nan 46.625347 51.58096 52.954323]
Resampled means: [51.780113 50.109566 50.84643 45.358025 63.376545 50.489815 50.38161
48.368156 48.42543 72.35379 47.457268 50.505123 83.48402 52.810825
48.268654 37.489105 46.84598 49.299854 50.877327 16.24648 47.929005
52.887367 48.82587 56.742073 55.441647 52.74524 46.725468 nan
28.190561 44.932938 43.253742 50.05856 47.255016 48.344154 49.795094
50.93613 nan 46.62006 51.485767 52.960278]
Chosen means: [51.780113 50.109566 50.84643 45.358025 63.376545 50.489815 50.38161
48.368156 48.42543 72.35379 47.457268 50.505123 83.48402 52.810825
48.268654 37.489105 46.84598 49.299854 50.877327 16.24648 47.929005
52.887367 48.82587 56.742073 55.441647 52.74524 46.725468 50.80916
28.190561 44.932938 43.253742 50.05856 47.255016 48.344154 49.795094
50.93613 51.840244 46.62006 51.485767 52.960278]
Category counts: [ 4 5 2 52 491 14 19 55 30 627 27 35 747 1 1 452 18 2
34 779 10 20 4 306 3 121 20 0 702 30 261 11 30 3 0 11
0 15 23 35]
New alpha: [4.312500e+00 5.312500e+00 2.312500e+00 5.231250e+01 4.913125e+02
1.431250e+01 1.931250e+01 5.531250e+01 3.031250e+01 6.273125e+02
2.731250e+01 3.531250e+01 7.473125e+02 1.312500e+00 1.312500e+00
4.523125e+02 1.831250e+01 2.312500e+00 3.431250e+01 7.793125e+02
1.031250e+01 2.031250e+01 4.312500e+00 3.063125e+02 3.312500e+00
1.213125e+02 2.031250e+01 3.125000e-01 7.023125e+02 3.031250e+01
2.613125e+02 1.131250e+01 3.031250e+01 3.312500e+00 3.125000e-01
1.131250e+01 3.125000e-01 1.531250e+01 2.331250e+01 3.531250e+01]
New probs: [5.50613215e-04 5.89958043e-04 8.51037330e-04 9.91903804e-03
9.98352915e-02 2.02510133e-03 3.71100078e-03 1.04941111e-02
5.26443031e-03 1.27247512e-01 6.88638166e-03 6.65142480e-03
1.51559144e-01 3.51287483e-04 8.47567047e-04 8.90266150e-02
2.78795138e-03 1.17548414e-04 7.49827968e-03 1.46313846e-01
2.54772091e-03 4.66085877e-03 1.20894052e-03 5.79411946e-02
1.02937956e-04 2.56634541e-02 3.61685851e-03 1.36729077e-05
1.44556224e-01 7.72151444e-03 5.51608168e-02 1.26656611e-03
7.85974879e-03 3.19221464e-04 1.77793263e-05 1.31565589e-03
2.93793692e-05 2.48040981e-03 4.59480612e-03 6.39364868e-03]
Category counts: [ 4 5 2 52 491 14 19 55 30 627 27 35 747 1 1 452 18 2
34 779 10 20 4 306 3 121 20 0 702 30 261 11 30 3 0 11
0 15 23 35]
Current means: [52.458218 50.369743 49.598595 45.142715 63.56568 50.46014 50.44936
48.34883 48.32494 72.52333 47.428776 50.45614 83.51126 52.271763
46.514797 36.867764 46.848923 49.310368 50.72709 15.749187 48.281227
53.45974 48.588776 57.154617 54.65057 53.09019 46.68873 nan
27.566729 44.97453 42.58965 49.95918 47.324566 48.262024 nan
50.279484 nan 46.603985 51.689762 53.486435]
Posterior variance: [2.4390244e-01 1.9607843e-01 4.7619051e-01 1.9193858e-02 2.0362451e-03
7.0921987e-02 5.2356020e-02 1.8148821e-02 3.3222590e-02 1.5946421e-03
3.6900368e-02 2.8490029e-02 1.3385089e-03 9.0909088e-01 9.0909088e-01
2.2119000e-03 5.5248618e-02 4.7619051e-01 2.9325515e-02 1.2835324e-03
9.9009894e-02 4.9751244e-02 2.4390244e-01 3.2669061e-03 3.2258067e-01
8.2576387e-03 4.9751244e-02 1.0000000e+01 1.4242986e-03 3.3222590e-02
3.8299502e-03 9.0090089e-02 3.3222590e-02 3.2258067e-01 1.0000000e+01
9.0090089e-02 1.0000000e+01 6.6225164e-02 4.3290041e-02 2.8490029e-02]
Posterior means: [52.398262 50.362495 49.617706 45.15204 63.562916 50.45688 50.44701
48.351826 48.330505 72.51974 47.438267 50.454838 83.506775 52.065243
46.83164 36.870667 46.866333 49.343204 50.72496 15.753584 48.29824
53.442528 48.6232 57.15228 54.500553 53.08764 46.7052 nan
27.569923 44.991226 42.592487 49.95955 47.333454 48.31809 nan
50.276966 nan 46.626476 51.68245 53.4765 ]
Resampled means: [52.099438 50.266426 49.798897 45.152996 63.565094 50.480465 50.335827
48.367916 48.29999 72.51891 47.430344 50.432423 83.50855 51.337162
46.489227 36.868164 46.75797 50.37275 50.78295 15.754236 48.296654
53.537384 48.525536 57.152035 54.414055 53.087955 46.70978 nan
27.569798 44.98782 42.58551 50.006172 47.352467 48.387543 nan
50.3885 nan 46.625797 51.681763 53.518127]
Chosen means: [52.099438 50.266426 49.798897 45.152996 63.565094 50.480465 50.335827
48.367916 48.29999 72.51891 47.430344 50.432423 83.50855 51.337162
46.489227 36.868164 46.75797 50.37275 50.78295 15.754236 48.296654
53.537384 48.525536 57.152035 54.414055 53.087955 46.70978 50.80916
27.569798 44.98782 42.58551 50.006172 47.352467 48.387543 49.795094
50.3885 51.840244 46.625797 51.681763 53.518127]
Category counts: [ 3 4 5 59 467 7 11 44 32 622 33 32 743 2 4 477 17 1
54 750 17 24 4 309 3 141 17 0 653 47 287 4 41 0 0 10
0 14 23 39]
New alpha: [3.312500e+00 4.312500e+00 5.312500e+00 5.931250e+01 4.673125e+02
7.312500e+00 1.131250e+01 4.431250e+01 3.231250e+01 6.223125e+02
3.331250e+01 3.231250e+01 7.433125e+02 2.312500e+00 4.312500e+00
4.773125e+02 1.731250e+01 1.312500e+00 5.431250e+01 7.503125e+02
1.731250e+01 2.431250e+01 4.312500e+00 3.093125e+02 3.312500e+00
1.413125e+02 1.731250e+01 3.125000e-01 6.533125e+02 4.731250e+01
2.873125e+02 4.312500e+00 4.131250e+01 3.125000e-01 3.125000e-01
1.031250e+01 3.125000e-01 1.431250e+01 2.331250e+01 3.931250e+01]
New probs: [3.95546667e-04 9.26361710e-04 6.23643631e-04 1.15436278e-02
8.69047493e-02 1.68144773e-03 3.25225107e-03 9.66177508e-03
6.37842622e-03 1.23308174e-01 6.28617359e-03 4.93882457e-03
1.48969725e-01 4.53699147e-04 8.27542390e-04 9.04025435e-02
4.04022262e-03 2.01229297e-04 1.05713550e-02 1.56989202e-01
2.83473823e-03 2.89690215e-03 9.57830227e-04 6.04429953e-02
8.31141486e-04 3.16211991e-02 3.16904671e-03 4.76798505e-06
1.33656248e-01 9.90589708e-03 5.73817268e-02 9.76520590e-04
9.30155534e-03 8.09296471e-05 6.41729184e-06 1.63199892e-03
1.21376498e-12 2.39990931e-03 5.21365786e-03 8.32996424e-03]
Category counts: [ 3 4 5 59 467 7 11 44 32 622 33 32 743 2 4 477 17 1
54 750 17 24 4 309 3 141 17 0 653 47 287 4 41 0 0 10
0 14 23 39]
Current means: [51.808483 50.997692 50.368397 45.03441 63.883102 50.744488 50.782078
48.555283 48.1733 72.62267 47.394314 50.538452 83.540375 51.659737
46.4226 35.772213 46.796494 48.478615 50.732597 15.501671 48.4006
54.03686 48.353374 57.72406 55.148346 53.23645 46.22799 nan
26.733555 45.195686 41.955124 50.59927 47.170322 nan nan
50.412052 nan 46.857746 51.302288 54.08124 ]
Posterior variance: [3.2258067e-01 2.4390244e-01 1.9607843e-01 1.6920473e-02 2.1408692e-03
1.4084508e-01 9.0090089e-02 2.2675738e-02 3.1152649e-02 1.6074587e-03
3.0211482e-02 3.1152649e-02 1.3457140e-03 4.7619051e-01 2.4390244e-01
2.0959966e-03 5.8479533e-02 9.0909088e-01 1.8484289e-02 1.3331557e-03
5.8479533e-02 4.1493773e-02 2.4390244e-01 3.2351988e-03 3.2258067e-01
7.0871720e-03 5.8479533e-02 1.0000000e+01 1.5311592e-03 2.1231424e-02
3.4831068e-03 2.4390244e-01 2.4330901e-02 1.0000000e+01 1.0000000e+01
9.9009894e-02 1.0000000e+01 7.0921987e-02 4.3290041e-02 2.5575448e-02]
Posterior means: [51.75015 50.973362 50.361176 45.042812 63.880135 50.734 50.775032
48.558563 48.178993 72.61903 47.402187 50.536774 83.535866 51.5807
46.509857 35.775196 46.815228 48.616924 50.731247 15.50627 48.409954
54.020107 48.39354 57.72156 54.982273 53.234154 46.250046 nan
26.737118 45.205887 41.957924 50.584656 47.177208 nan nan
50.40797 nan 46.88003 51.296654 54.0708 ]
Resampled means: [51.969234 50.99517 50.222664 45.059414 63.878696 50.842064 50.73667
48.57162 48.23932 72.61995 47.411255 50.578503 83.5339 52.45607
46.7497 35.773148 46.834297 48.45116 50.774162 15.506861 48.39413
54.0277 48.4398 57.72023 54.82436 53.227547 46.27574 nan
26.740126 45.205597 41.96135 50.62169 47.139923 nan nan
50.59326 nan 46.82869 51.222286 54.024208]
Chosen means: [51.969234 50.99517 50.222664 45.059414 63.878696 50.842064 50.73667
48.57162 48.23932 72.61995 47.411255 50.578503 83.5339 52.45607
46.7497 35.773148 46.834297 48.45116 50.774162 15.506861 48.39413
54.0277 48.4398 57.72023 54.82436 53.227547 46.27574 50.80916
26.740126 45.205597 41.96135 50.62169 47.139923 48.387543 49.795094
50.59326 51.840244 46.82869 51.222286 54.024208]
Category counts: [ 3 4 2 69 433 8 15 45 29 608 30 32 741 4 3 478 25 1
51 747 11 13 4 341 3 156 7 0 628 58 292 5 52 1 0 6
0 13 33 49]
New alpha: [3.312500e+00 4.312500e+00 2.312500e+00 6.931250e+01 4.333125e+02
8.312500e+00 1.531250e+01 4.531250e+01 2.931250e+01 6.083125e+02
3.031250e+01 3.231250e+01 7.413125e+02 4.312500e+00 3.312500e+00
4.783125e+02 2.531250e+01 1.312500e+00 5.131250e+01 7.473125e+02
1.131250e+01 1.331250e+01 4.312500e+00 3.413125e+02 3.312500e+00
1.563125e+02 7.312500e+00 3.125000e-01 6.283125e+02 5.831250e+01
2.923125e+02 5.312500e+00 5.231250e+01 1.312500e+00 3.125000e-01
6.312500e+00 3.125000e-01 1.331250e+01 3.331250e+01 4.931250e+01]
New probs: [4.57213959e-04 5.41546324e-04 4.11632092e-04 1.27824871e-02
8.97891894e-02 2.51931092e-03 2.93767359e-03 1.05638672e-02
8.19684379e-03 1.14445135e-01 4.11683740e-03 7.74668111e-03
1.42698258e-01 9.43477033e-04 2.25406548e-04 9.87161249e-02
7.52458395e-03 1.68234576e-04 1.35036502e-02 1.53152034e-01
2.16076011e-03 3.59008787e-03 7.48401624e-04 6.67696744e-02
8.28044489e-04 3.30540650e-02 1.26991700e-03 1.72285564e-08
1.16641164e-01 1.14026573e-02 6.11879155e-02 9.55115655e-04
1.03240283e-02 2.85747141e-04 4.46512473e-07 1.65055471e-03
7.74601504e-05 2.56015710e-03 7.06391782e-03 7.98929110e-03]
Category counts: [ 3 4 2 69 433 8 15 45 29 608 30 32 741 4 3 478 25 1
51 747 11 13 4 341 3 156 7 0 628 58 292 5 52 1 0 6
0 13 33 49]
Current means: [52.161163 50.50918 50.46067 45.043655 64.423775 50.294605 50.771114
48.55644 48.194923 72.75807 46.91361 50.48223 83.554405 52.74115
46.751465 35.37449 46.78631 48.44412 50.624996 15.479738 48.591366
54.29083 48.23803 58.290546 55.243702 53.347717 46.11414 nan
26.464443 44.959396 41.494717 49.89142 47.185043 48.098957 nan
51.029373 nan 46.937664 51.519917 54.182514]
Posterior variance: [3.22580665e-01 2.43902445e-01 4.76190507e-01 1.44717805e-02
2.30893563e-03 1.23456784e-01 6.62251636e-02 2.21729502e-02
3.43642607e-02 1.64446642e-03 3.32225896e-02 3.11526489e-02
1.34934566e-03 2.43902445e-01 3.22580665e-01 2.09161267e-03
3.98406386e-02 9.09090877e-01 1.95694715e-02 1.33850891e-03
9.00900885e-02 7.63358772e-02 2.43902445e-01 2.93169147e-03
3.22580665e-01 6.40614983e-03 1.40845075e-01 1.00000000e+01
1.59210328e-03 1.72117036e-02 3.42348497e-03 1.96078435e-01
1.91938579e-02 9.09090877e-01 1.00000000e+01 1.63934425e-01
1.00000000e+01 7.63358772e-02 3.02114822e-02 2.03665998e-02]
Posterior means: [52.09145 50.49676 50.438732 45.050827 64.42044 50.290966 50.76601
48.55964 48.201126 72.754326 46.923862 50.48073 83.54987 52.674294
46.85626 35.37755 46.799114 48.585564 50.62377 15.484359 48.604057
54.25807 48.281006 58.288116 55.07455 53.345573 46.16887 nan
26.46819 44.96807 41.497627 49.893547 47.19045 48.271782 nan
51.012497 nan 46.96104 51.515327 54.173996]
Resampled means: [51.778454 50.65812 50.437397 45.033436 64.419525 50.30869 50.712044
48.564934 48.217964 72.75213 46.95244 50.435547 83.55037 52.371468
46.870842 35.3771 46.730667 47.509846 50.64816 15.484293 48.572617
54.29923 48.338905 58.29456 54.61717 53.350925 46.106586 nan
26.468958 44.99735 41.498302 50.016335 47.200733 47.369595 nan
51.223503 nan 47.019794 51.448093 54.220955]
Chosen means: [51.778454 50.65812 50.437397 45.033436 64.419525 50.30869 50.712044
48.564934 48.217964 72.75213 46.95244 50.435547 83.55037 52.371468
46.870842 35.3771 46.730667 47.509846 50.64816 15.484293 48.572617
54.29923 48.338905 58.29456 54.61717 53.350925 46.106586 50.80916
26.468958 44.99735 41.498302 50.016335 47.200733 47.369595 49.795094
51.223503 51.840244 47.019794 51.448093 54.220955]
Category counts: [ 2 1 3 89 451 11 13 52 24 570 17 35 731 1 0 448 34 2
61 739 15 20 6 358 5 159 2 0 621 62 317 3 48 1 0 13
0 15 29 42]
New alpha: [2.312500e+00 1.312500e+00 3.312500e+00 8.931250e+01 4.513125e+02
1.131250e+01 1.331250e+01 5.231250e+01 2.431250e+01 5.703125e+02
1.731250e+01 3.531250e+01 7.313125e+02 1.312500e+00 3.125000e-01
4.483125e+02 3.431250e+01 2.312500e+00 6.131250e+01 7.393125e+02
1.531250e+01 2.031250e+01 6.312500e+00 3.583125e+02 5.312500e+00
1.593125e+02 2.312500e+00 3.125000e-01 6.213125e+02 6.231250e+01
3.173125e+02 3.312500e+00 4.831250e+01 1.312500e+00 3.125000e-01
1.331250e+01 3.125000e-01 1.531250e+01 2.931250e+01 4.231250e+01]
New probs: [1.17521174e-03 9.74987488e-05 3.86229483e-04 1.83930509e-02
8.88599828e-02 3.79774044e-03 1.88979856e-03 1.01459995e-02
3.01949191e-03 1.13226585e-01 1.84643606e-03 4.76751011e-03
1.52285963e-01 3.81575985e-04 2.27941865e-09 9.11459923e-02
6.20168122e-03 3.81296442e-04 1.18881203e-02 1.42691597e-01
3.32214939e-03 3.50082526e-03 1.78895739e-03 7.69537315e-02
8.03280389e-04 2.89097130e-02 4.63055214e-04 6.90974121e-05
1.21075727e-01 1.27112456e-02 6.50436282e-02 4.65485093e-04
1.08919116e-02 1.11119734e-04 1.52141647e-05 2.85429182e-03
3.30163493e-06 2.37085391e-03 8.94798711e-03 7.11640203e-03]
Category counts: [ 2 1 3 89 451 11 13 52 24 570 17 35 731 1 0 448 34 2
61 739 15 20 6 358 5 159 2 0 621 62 317 3 48 1 0 13
0 15 29 42]
Current means: [52.38581 52.27214 50.714767 44.949074 65.0924 50.13946 50.13482
48.646824 48.38838 73.21808 46.83548 50.382942 83.628204 52.126312
nan 34.89329 46.15352 47.552677 50.538105 15.421519 48.76348
54.600525 48.31547 58.589172 54.54917 53.504574 47.552326 nan
26.286446 44.83308 40.938133 50.626232 47.288136 48.173595 nan
51.35458 nan 47.025867 51.65054 54.63077 ]
Posterior variance: [4.7619051e-01 9.0909088e-01 3.2258067e-01 1.1223345e-02 2.2168034e-03
9.0090089e-02 7.6335877e-02 1.9193858e-02 4.1493773e-02 1.7540783e-03
5.8479533e-02 2.8490029e-02 1.3678020e-03 9.0909088e-01 1.0000000e+01
2.2316447e-03 2.9325515e-02 4.7619051e-01 1.6366612e-02 1.3529969e-03
6.6225164e-02 4.9751244e-02 1.6393442e-01 2.7925160e-03 1.9607843e-01
6.2853550e-03 4.7619051e-01 1.0000000e+01 1.6100467e-03 1.6103059e-02
3.1535793e-03 3.2258067e-01 2.0790022e-02 9.0909088e-01 1.0000000e+01
7.6335877e-02 1.0000000e+01 6.6225164e-02 3.4364261e-02 2.3752971e-02]
Posterior means: [52.272198 52.065586 50.69171 44.954742 65.08906 50.138206 50.13379
48.64942 48.395065 73.21401 46.853985 50.38185 83.623604 51.933014
nan 34.89666 46.164803 47.669216 50.537224 15.426198 48.77167
54.577637 48.343086 58.586773 54.459972 53.50237 47.66888 nan
26.290262 44.8414 40.94099 50.606033 47.293777 48.339634 nan
51.34424 nan 47.045567 51.644867 54.61977 ]
Resampled means: [52.46595 53.024315 50.709507 44.94504 65.09227 50.146793 50.100792
48.611645 48.424084 73.214554 46.81184 50.38269 83.625916 53.336548
nan 34.89911 46.14252 48.424255 50.554794 15.426304 48.847336
54.54357 48.297623 58.592747 54.19883 53.49225 46.791092 nan
26.287336 44.84804 40.938633 50.177395 47.31273 48.776417 nan
51.28399 nan 46.95205 51.62748 54.5733 ]
Chosen means: [52.46595 53.024315 50.709507 44.94504 65.09227 50.146793 50.100792
48.611645 48.424084 73.214554 46.81184 50.38269 83.625916 53.336548
46.870842 34.89911 46.14252 48.424255 50.554794 15.426304 48.847336
54.54357 48.297623 58.592747 54.19883 53.49225 46.791092 50.80916
26.287336 44.84804 40.938633 50.177395 47.31273 48.776417 49.795094
51.28399 51.840244 46.95205 51.62748 54.5733 ]
Category counts: [ 7 0 0 95 484 12 10 52 21 563 8 25 696 1 0 436 26 2
53 730 19 28 3 348 6 143 2 1 595 73 354 1 59 2 0 22
0 13 52 58]
New alpha: [7.312500e+00 3.125000e-01 3.125000e-01 9.531250e+01 4.843125e+02
1.231250e+01 1.031250e+01 5.231250e+01 2.131250e+01 5.633125e+02
8.312500e+00 2.531250e+01 6.963125e+02 1.312500e+00 3.125000e-01
4.363125e+02 2.631250e+01 2.312500e+00 5.331250e+01 7.303125e+02
1.931250e+01 2.831250e+01 3.312500e+00 3.483125e+02 6.312500e+00
1.433125e+02 2.312500e+00 1.312500e+00 5.953125e+02 7.331250e+01
3.543125e+02 1.312500e+00 5.931250e+01 2.312500e+00 3.125000e-01
2.231250e+01 3.125000e-01 1.331250e+01 5.231250e+01 5.831250e+01]
New probs: [5.53623424e-04 6.71854941e-05 1.73792650e-05 1.93275586e-02
9.73945335e-02 3.25491326e-03 1.93246093e-03 9.41995811e-03
5.38600516e-03 1.09348126e-01 1.92798686e-03 4.18385025e-03
1.39381498e-01 6.73603499e-04 1.76259655e-05 8.77400115e-02
5.63173089e-03 8.32784688e-04 1.03981262e-02 1.53623939e-01
2.45416258e-03 6.38507539e-03 1.40874658e-03 6.30565509e-02
9.02390049e-04 2.74923816e-02 5.74995007e-04 2.43396469e-04
1.16729468e-01 1.41512649e-02 7.22115859e-02 1.14792885e-04
1.19977798e-02 2.20109214e-04 1.35121634e-04 3.84131633e-03
6.66408084e-08 2.60680774e-03 1.13198133e-02 1.30412579e-02]
Category counts: [ 7 0 0 95 484 12 10 52 21 563 8 25 696 1 0 436 26 2
53 730 19 28 3 348 6 143 2 1 595 73 354 1 59 2 0 22
0 13 52 58]
Current means: [52.487514 nan nan 44.91776 65.468155 49.81493 49.814453
48.632366 48.05768 73.86622 46.561493 50.250607 83.898445 54.448006
nan 34.181686 46.345238 49.323444 50.53909 15.354408 48.85719
55.004173 48.00366 58.800224 54.314503 53.607487 46.35243 48.378395
25.944597 44.689327 40.527004 50.96304 47.370476 47.75454 nan
51.412785 nan 46.878723 51.295147 54.807343]
Posterior variance: [1.40845075e-01 1.00000000e+01 1.00000000e+01 1.05152475e-02
2.06568884e-03 8.26446265e-02 9.90098938e-02 1.91938579e-02
4.73933630e-02 1.77588360e-03 1.23456784e-01 3.98406386e-02
1.43657520e-03 9.09090877e-01 1.00000000e+01 2.29305192e-03
3.83141749e-02 4.76190507e-01 1.88323930e-02 1.36967539e-03
5.23560196e-02 3.55871879e-02 3.22580665e-01 2.87273759e-03
1.63934425e-01 6.98811980e-03 4.76190507e-01 9.09090877e-01
1.68038986e-03 1.36798909e-02 2.82406085e-03 9.09090877e-01
1.69204734e-02 4.76190507e-01 1.00000000e+01 4.52488698e-02
1.00000000e+01 7.63358772e-02 1.91938579e-02 1.72117036e-02]
Posterior means: [52.452477 nan nan 44.923103 65.464966 49.816463 49.816288
48.63499 48.066883 73.861984 46.603943 50.24961 83.89357 54.043644
nan 34.185314 46.359238 49.35566 50.538074 15.359155 48.863174
54.986366 48.068058 58.7977 54.24377 53.60496 46.52612 48.525818
25.948639 44.696594 40.529682 50.875492 47.374928 47.861465 nan
51.40639 nan 46.90255 51.292664 54.79907 ]
Resampled means: [52.364983 nan nan 44.919765 65.46381 49.857327 49.7181
48.61851 48.143475 73.862144 46.663307 50.25166 83.896034 55.85802
nan 34.18826 46.304985 50.47348 50.5274 15.358143 48.86103
54.972214 47.72297 58.79662 54.369667 53.60536 46.772152 49.767803
25.946922 44.70841 40.528233 51.804516 47.37206 48.25525 nan
51.420113 nan 47.027935 51.312702 54.82441 ]
Chosen means: [52.364983 53.024315 50.709507 44.919765 65.46381 49.857327 49.7181
48.61851 48.143475 73.862144 46.663307 50.25166 83.896034 55.85802
46.870842 34.18826 46.304985 50.47348 50.5274 15.358143 48.86103
54.972214 47.72297 58.79662 54.369667 53.60536 46.772152 49.767803
25.946922 44.70841 40.528233 51.804516 47.37206 48.25525 49.795094
51.420113 51.840244 47.027935 51.312702 54.82441 ]
Category counts: [ 4 1 0 98 486 16 11 43 36 607 15 12 641 9 1 473 25 11
54 705 11 43 10 322 4 141 2 0 570 93 343 1 49 0 0 19
0 14 54 76]
New alpha: [4.312500e+00 1.312500e+00 3.125000e-01 9.831250e+01 4.863125e+02
1.631250e+01 1.131250e+01 4.331250e+01 3.631250e+01 6.073125e+02
1.531250e+01 1.231250e+01 6.413125e+02 9.312500e+00 1.312500e+00
4.733125e+02 2.531250e+01 1.131250e+01 5.431250e+01 7.053125e+02
1.131250e+01 4.331250e+01 1.031250e+01 3.223125e+02 4.312500e+00
1.413125e+02 2.312500e+00 3.125000e-01 5.703125e+02 9.331250e+01
3.433125e+02 1.312500e+00 4.931250e+01 3.125000e-01 3.125000e-01
1.931250e+01 3.125000e-01 1.431250e+01 5.431250e+01 7.631250e+01]
New probs: [5.19645284e-04 1.28823347e-04 7.48874936e-06 2.08237488e-02
9.94206518e-02 2.33354419e-03 2.61380291e-03 9.26231872e-03
6.12888299e-03 1.25395983e-01 2.77870358e-03 2.98502971e-03
1.23296998e-01 2.00855639e-03 4.91544488e-04 9.53863189e-02
4.67956299e-03 1.78524887e-03 1.16895176e-02 1.37271509e-01
1.44212251e-03 9.72309615e-03 1.84434617e-03 6.75229207e-02
5.26134623e-04 3.24342512e-02 4.40413889e-04 9.95115546e-09
1.07956588e-01 1.93340573e-02 6.81864992e-02 4.67171449e-05
9.30889416e-03 4.59534058e-05 2.87953808e-05 3.74271837e-03
3.68633046e-05 3.12265079e-03 9.65835247e-03 1.55903855e-02]
Category counts: [ 4 1 0 98 486 16 11 43 36 607 15 12 641 9 1 473 25 11
54 705 11 43 10 322 4 141 2 0 570 93 343 1 49 0 0 19
0 14 54 76]
Current means: [51.95565 52.086338 nan 44.83455 65.61099 49.757565 50.1285
48.47998 48.110405 74.38119 46.54004 50.133915 84.3537 55.878277
46.820255 33.673515 46.424423 50.72029 50.54253 15.161133 49.036343
55.30679 47.314346 59.146812 54.542953 53.54481 46.910088 nan
25.330671 44.320442 40.270836 52.43161 47.429276 nan nan
51.64335 nan 46.801167 51.307793 55.28075 ]
Posterior variance: [2.4390244e-01 9.0909088e-01 1.0000000e+01 1.0193680e-02 2.0571898e-03
6.2111799e-02 9.0090089e-02 2.3201857e-02 2.7700832e-02 1.6471752e-03
6.6225164e-02 8.2644626e-02 1.5598191e-03 1.0989010e-01 9.0909088e-01
2.1137181e-03 3.9840639e-02 9.0090089e-02 1.8484289e-02 1.4182386e-03
9.0090089e-02 2.3201857e-02 9.9009894e-02 3.1046257e-03 2.4390244e-01
7.0871720e-03 4.7619051e-01 1.0000000e+01 1.7540783e-03 1.0741139e-02
2.9146022e-03 9.0909088e-01 2.0366600e-02 1.0000000e+01 1.0000000e+01
5.2356020e-02 1.0000000e+01 7.0921987e-02 1.8484289e-02 1.3140605e-02]
Posterior means: [51.907955 51.896675 nan 44.839813 65.60778 49.75907 50.127346
48.48351 48.11564 74.37717 46.562954 50.13281 84.348335 55.813683
47.109325 33.676964 46.43867 50.713802 50.54153 15.166075 49.045025
55.29448 47.340935 59.143974 54.43215 53.542297 47.057224 nan
25.334997 44.32654 40.27367 52.210556 47.43451 nan nan
51.634747 nan 46.823856 51.305378 55.273815]
Resampled means: [52.22246 51.18682 nan 44.82637 65.60758 49.796207 50.162277
48.50007 48.12465 74.37893 46.541866 49.945393 84.34856 55.73138
47.16697 33.6751 46.47154 50.671124 50.524315 15.166265 48.9951
55.313812 47.232914 59.14155 54.58761 53.544086 47.78888 nan
25.338017 44.32581 40.27361 51.50878 47.410873 nan nan
51.67362 nan 46.970184 51.327717 55.285606]
Chosen means: [52.22246 51.18682 50.709507 44.82637 65.60758 49.796207 50.162277
48.50007 48.12465 74.37893 46.541866 49.945393 84.34856 55.73138
47.16697 33.6751 46.47154 50.671124 50.524315 15.166265 48.9951
55.313812 47.232914 59.14155 54.58761 53.544086 47.78888 49.767803
25.338017 44.32581 40.27361 51.50878 47.410873 48.25525 49.795094
51.67362 51.840244 46.970184 51.327717 55.285606]
Category counts: [ 3 0 0 105 472 6 12 42 35 614 13 11 626 14 3 483 24 10
57 663 10 58 13 307 4 137 7 0 588 111 333 1 45 0 0 23
0 12 69 89]
New alpha: [3.312500e+00 3.125000e-01 3.125000e-01 1.053125e+02 4.723125e+02
6.312500e+00 1.231250e+01 4.231250e+01 3.531250e+01 6.143125e+02
1.331250e+01 1.131250e+01 6.263125e+02 1.431250e+01 3.312500e+00
4.833125e+02 2.431250e+01 1.031250e+01 5.731250e+01 6.633125e+02
1.031250e+01 5.831250e+01 1.331250e+01 3.073125e+02 4.312500e+00
1.373125e+02 7.312500e+00 3.125000e-01 5.883125e+02 1.113125e+02
3.333125e+02 1.312500e+00 4.531250e+01 3.125000e-01 3.125000e-01
2.331250e+01 3.125000e-01 1.231250e+01 6.931250e+01 8.931250e+01]
New probs: [1.0126240e-03 1.7629663e-05 1.7853152e-05 1.9094646e-02 9.5397457e-02
8.6730911e-04 2.9868223e-03 9.3878917e-03 7.0895334e-03 1.1876182e-01
1.8820737e-03 2.1900993e-03 1.3107361e-01 3.5304704e-03 2.7596194e-04
9.7001828e-02 5.2733496e-03 1.0589143e-03 1.4671464e-02 1.2943423e-01
2.1019971e-03 1.1797468e-02 2.7264454e-03 6.0493395e-02 4.6705676e-04
2.9943962e-02 1.2871934e-03 6.6768515e-09 1.2130058e-01 1.8849092e-02
5.9338003e-02 3.7183959e-04 8.3698193e-03 1.5454173e-04 9.5853429e-05
6.0368939e-03 5.0103813e-06 2.2793300e-03 1.5063850e-02 1.8292073e-02]
Category counts: [ 3 0 0 105 472 6 12 42 35 614 13 11 626 14 3 483 24 10
57 663 10 58 13 307 4 137 7 0 588 111 333 1 45 0 0 23
0 12 69 89]
Current means: [52.658916 nan nan 44.609146 65.84088 49.65028 50.115845
48.345806 48.27264 74.5459 46.53332 50.108604 84.486084 56.141327
46.61653 33.404823 46.67662 50.570908 50.57959 14.824567 49.197697
55.81433 47.331577 59.66999 54.10599 53.468185 48.02641 nan
24.788593 44.046066 39.96754 52.102142 47.376987 nan nan
51.507008 nan 47.008137 51.203453 55.747772]
Posterior variance: [3.22580665e-01 1.00000000e+01 1.00000000e+01 9.51474812e-03
2.11819517e-03 1.63934425e-01 8.26446265e-02 2.37529706e-02
2.84900293e-02 1.62839936e-03 7.63358772e-02 9.00900885e-02
1.59718900e-03 7.09219873e-02 3.22580665e-01 2.06996477e-03
4.14937735e-02 9.90098938e-02 1.75131354e-02 1.50806818e-03
9.90098938e-02 1.72117036e-02 7.63358772e-02 3.25626833e-03
2.43902445e-01 7.29394564e-03 1.40845075e-01 1.00000000e+01
1.70039118e-03 9.00090020e-03 3.00210132e-03 9.09090877e-01
2.21729502e-02 1.00000000e+01 1.00000000e+01 4.32900414e-02
1.00000000e+01 8.26446265e-02 1.44717805e-02 1.12233451e-02]
Posterior means: [52.573147 nan nan 44.614277 65.837524 49.65601 50.11489
48.34973 48.27756 74.5419 46.559784 50.107628 84.480576 56.097775
46.725677 33.408257 46.69041 50.56525 50.578575 14.829871 49.205635
55.80432 47.351944 59.66684 54.005848 53.465656 48.054203 nan
24.792881 44.051426 39.97055 51.91104 47.3828 nan nan
51.500484 nan 47.032864 51.20171 55.74132 ]
Resampled means: [52.879013 nan nan 44.612263 65.83939 49.7814 50.245808
48.324745 48.268925 74.54091 46.634705 50.156254 84.47979 56.16387
47.06816 33.407253 46.70108 50.470577 50.591915 14.832874 49.046524
55.821815 47.34488 59.66528 54.24615 53.463367 47.908268 nan
24.793262 44.04193 39.974266 50.108757 47.417953 nan nan
51.41782 nan 47.022514 51.216152 55.74067 ]
Chosen means: [52.879013 51.18682 50.709507 44.612263 65.83939 49.7814 50.245808
48.324745 48.268925 74.54091 46.634705 50.156254 84.47979 56.16387
47.06816 33.407253 46.70108 50.470577 50.591915 14.832874 49.046524
55.821815 47.34488 59.66528 54.24615 53.463367 47.908268 49.767803
24.793262 44.04193 39.974266 50.108757 47.417953 48.25525 49.795094
51.41782 51.840244 47.022514 51.216152 55.74067 ]
Category counts: [ 6 0 0 103 441 4 8 58 42 599 10 11 625 21 3 467 26 4
63 626 8 62 19 335 0 136 5 0 622 123 326 2 46 0 0 13
0 12 74 100]
New alpha: [6.312500e+00 3.125000e-01 3.125000e-01 1.033125e+02 4.413125e+02
4.312500e+00 8.312500e+00 5.831250e+01 4.231250e+01 5.993125e+02
1.031250e+01 1.131250e+01 6.253125e+02 2.131250e+01 3.312500e+00
4.673125e+02 2.631250e+01 4.312500e+00 6.331250e+01 6.263125e+02
8.312500e+00 6.231250e+01 1.931250e+01 3.353125e+02 3.125000e-01
1.363125e+02 5.312500e+00 3.125000e-01 6.223125e+02 1.233125e+02
3.263125e+02 2.312500e+00 4.631250e+01 3.125000e-01 3.125000e-01
1.331250e+01 3.125000e-01 1.231250e+01 7.431250e+01 1.003125e+02]
New probs: [6.30164461e-04 4.21892473e-05 2.07204766e-06 2.00168956e-02
9.28333551e-02 3.41175270e-04 2.37114774e-03 1.18548917e-02
9.50117689e-03 1.31389424e-01 1.18149142e-03 2.72683287e-03
1.11796945e-01 4.94735688e-03 7.79976195e-04 8.91217142e-02
5.31393848e-03 1.72602653e-04 1.47178508e-02 1.20143764e-01
2.17816606e-03 1.21191246e-02 3.99389304e-03 6.62586614e-02
5.94516166e-07 2.95022782e-02 7.22188153e-04 1.31197958e-05
1.20642178e-01 2.55010072e-02 6.73852563e-02 2.75583734e-04
1.05297901e-02 1.07815567e-05 1.03960083e-05 1.47636991e-03
3.65277515e-06 3.43346363e-03 1.85732134e-02 1.74854230e-02]
Category counts: [ 6 0 0 103 441 4 8 58 42 599 10 11 625 21 3 467 26 4
63 626 8 62 19 335 0 136 5 0 622 123 326 2 46 0 0 13
0 12 74 100]
Current means: [52.607983 nan nan 44.394646 66.357635 49.88905 50.33603
48.266964 48.139874 74.669205 46.86089 50.128414 84.49509 55.96239
46.498306 33.24319 46.34943 51.655327 50.589874 14.506864 49.26416
55.891277 47.175262 60.216652 nan 53.540222 47.56932 nan
24.494093 43.655155 39.575134 51.101486 47.29506 nan nan
51.53151 nan 46.78593 51.31214 55.8261 ]
Posterior variance: [1.6393442e-01 1.0000000e+01 1.0000000e+01 9.6993214e-03 2.2670596e-03
2.4390244e-01 1.2345678e-01 1.7211704e-02 2.3752971e-02 1.6691705e-03
9.9009894e-02 9.0090089e-02 1.5997441e-03 4.7393363e-02 3.2258067e-01
2.1408692e-03 3.8314175e-02 2.4390244e-01 1.5847862e-02 1.5971890e-03
1.2345678e-01 1.6103059e-02 5.2356020e-02 2.9841838e-03 1.0000000e+01
7.3475381e-03 1.9607843e-01 1.0000000e+01 1.6074587e-03 8.1234770e-03
3.0665440e-03 4.7619051e-01 2.1691974e-02 1.0000000e+01 1.0000000e+01
7.6335877e-02 1.0000000e+01 8.2644626e-02 1.3495277e-02 9.9900104e-03]
Posterior means: [52.565228 nan nan 44.400085 66.35393 49.891758 50.33188
48.269947 48.14429 74.665085 46.891968 50.12726 84.48956 55.93413
46.611267 33.24678 46.363415 51.614956 50.58894 14.512532 49.273243
55.88179 47.190052 60.2136 nan 53.537624 47.61698 nan
24.498194 43.66031 39.57833 51.049034 47.300926 nan nan
51.519817 nan 46.812496 51.310375 55.82028 ]
Resampled means: [52.475998 nan nan 44.394966 66.34895 49.64899 50.535816
48.23242 48.173904 74.66457 46.949863 50.076054 84.48984 55.86026
46.43058 33.249573 46.389717 51.80691 50.58597 14.512575 49.203236
55.896664 47.237 60.212715 nan 53.52616 47.624783 nan
24.498447 43.649147 39.580177 50.64512 47.2651 nan nan
51.488792 nan 46.8697 51.2964 55.834988]
Chosen means: [52.475998 51.18682 50.709507 44.394966 66.34895 49.64899 50.535816
48.23242 48.173904 74.66457 46.949863 50.076054 84.48984 55.86026
46.43058 33.249573 46.389717 51.80691 50.58597 14.512575 49.203236
55.896664 47.237 60.212715 54.24615 53.52616 47.624783 49.767803
24.498447 43.649147 39.580177 50.64512 47.2651 48.25525 49.795094
51.488792 51.840244 46.8697 51.2964 55.834988]
Category counts: [ 3 0 0 89 438 1 5 58 41 567 3 8 625 29 6 445 25 1
67 625 14 64 19 354 0 143 4 0 605 151 334 1 60 0 0 5
0 21 98 91]
New alpha: [3.312500e+00 3.125000e-01 3.125000e-01 8.931250e+01 4.383125e+02
1.312500e+00 5.312500e+00 5.831250e+01 4.131250e+01 5.673125e+02
3.312500e+00 8.312500e+00 6.253125e+02 2.931250e+01 6.312500e+00
4.453125e+02 2.531250e+01 1.312500e+00 6.731250e+01 6.253125e+02
1.431250e+01 6.431250e+01 1.931250e+01 3.543125e+02 3.125000e-01
1.433125e+02 4.312500e+00 3.125000e-01 6.053125e+02 1.513125e+02
3.343125e+02 1.312500e+00 6.031250e+01 3.125000e-01 3.125000e-01
5.312500e+00 3.125000e-01 2.131250e+01 9.831250e+01 9.131250e+01]
New probs: [2.11271385e-04 1.35838782e-05 1.27796957e-04 1.72586627e-02
8.72449651e-02 1.00530677e-04 8.74617661e-04 1.00507680e-02
8.78628064e-03 1.06764004e-01 1.16220664e-03 1.86295179e-03
1.28985152e-01 6.08831085e-03 1.51184166e-03 8.56430903e-02
5.16845938e-03 2.78661428e-05 1.49070835e-02 1.31678000e-01
2.93801539e-03 1.15314191e-02 3.04642506e-03 6.82654828e-02
3.05764661e-05 2.79406942e-02 5.56102546e-04 6.20915216e-07
1.27709493e-01 2.58768499e-02 6.25157654e-02 7.57156406e-04
1.23303346e-02 7.23897174e-05 1.20159850e-04 8.93591554e-04
1.18760967e-04 4.11548698e-03 2.15926152e-02 2.11207438e-02]
Category counts: [ 3 0 0 89 438 1 5 58 41 567 3 8 625 29 6 445 25 1
67 625 14 64 19 354 0 143 4 0 605 151 334 1 60 0 0 5
0 21 98 91]
Current means: [53.368908 nan nan 44.185307 66.94094 50.13239
50.627583 48.22239 48.208336 74.91104 46.747406 49.34056
84.49509 56.13311 46.897816 32.761677 46.63569 52.22962
50.429424 14.4979315 50.087307 56.076782 47.2142 60.574764
nan 53.596214 47.322098 nan 24.358988 43.243004
38.959534 50.72257 47.107735 nan nan 50.92023
nan 46.302166 51.55475 56.257233 ]
Posterior variance: [3.22580665e-01 1.00000000e+01 1.00000000e+01 1.12233451e-02
2.28258385e-03 9.09090877e-01 1.96078435e-01 1.72117036e-02
2.43309010e-02 1.76335755e-03 3.22580665e-01 1.23456784e-01
1.59974408e-03 3.43642607e-02 1.63934425e-01 2.24668602e-03
3.98406386e-02 9.09090877e-01 1.49031300e-02 1.59974408e-03
7.09219873e-02 1.56006245e-02 5.23560196e-02 2.82406085e-03
1.00000000e+01 6.98811980e-03 2.43902445e-01 1.00000000e+01
1.65261945e-03 6.61813328e-03 2.99311569e-03 9.09090877e-01
1.66389346e-02 1.00000000e+01 1.00000000e+01 1.96078435e-01
1.00000000e+01 4.73933630e-02 1.01936804e-02 1.09769488e-02]
Posterior means: [53.260235 nan nan 44.191833 66.93707 50.120358 50.615276
48.22545 48.212696 74.906654 46.85233 49.3487 84.48956 56.112034
46.94867 32.765553 46.649094 52.02693 50.428787 14.503611 50.08669
56.0673 47.228786 60.57178 nan 53.593697 47.387413 nan
24.363226 43.24748 38.96284 50.656883 47.11255 nan nan
50.902187 nan 46.31969 51.55316 56.250362]
Resampled means: [52.962532 nan nan 44.182682 66.934 50.01998 50.56685
48.21082 48.196022 74.90759 47.13692 49.287994 84.49061 56.082504
46.839577 32.767048 46.7041 50.835552 50.420532 14.50062 50.015728
56.10189 47.256382 60.567444 nan 53.600033 47.894764 nan
24.365772 43.26063 38.963196 51.431324 47.08353 nan nan
50.601242 nan 46.34021 51.536587 56.2393 ]
Chosen means: [52.962532 51.18682 50.709507 44.182682 66.934 50.01998 50.56685
48.21082 48.196022 74.90759 47.13692 49.287994 84.49061 56.082504
46.839577 32.767048 46.7041 50.835552 50.420532 14.50062 50.015728
56.10189 47.256382 60.567444 54.24615 53.600033 47.894764 49.767803
24.365772 43.26063 38.963196 51.431324 47.08353 48.25525 49.795094
50.601242 51.840244 46.34021 51.536587 56.2393 ]
Category counts: [ 0 0 1 93 474 0 7 40 44 526 9 10 623 32 9 432 29 0
68 626 22 53 14 337 0 137 6 0 574 165 360 5 59 0 1 2
0 19 101 122]
New alpha: [3.125000e-01 3.125000e-01 1.312500e+00 9.331250e+01 4.743125e+02
3.125000e-01 7.312500e+00 4.031250e+01 4.431250e+01 5.263125e+02
9.312500e+00 1.031250e+01 6.233125e+02 3.231250e+01 9.312500e+00
4.323125e+02 2.931250e+01 3.125000e-01 6.831250e+01 6.263125e+02
2.231250e+01 5.331250e+01 1.431250e+01 3.373125e+02 3.125000e-01
1.373125e+02 6.312500e+00 3.125000e-01 5.743125e+02 1.653125e+02
3.603125e+02 5.312500e+00 5.931250e+01 3.125000e-01 1.312500e+00
2.312500e+00 3.125000e-01 1.931250e+01 1.013125e+02 1.223125e+02]
New probs: [1.13273654e-06 5.27891934e-05 6.20529827e-05 1.98495425e-02
9.72557217e-02 7.68149960e-07 6.45295426e-04 6.90966612e-03
1.05500240e-02 1.06586933e-01 2.42289808e-03 2.32456834e-03
1.21450998e-01 8.18770658e-03 1.80907350e-03 8.88506621e-02
6.34521805e-03 4.03653848e-06 1.36121241e-02 1.22314282e-01
4.54292959e-03 9.76609066e-03 2.06454960e-03 6.91548064e-02
1.76261587e-04 2.79934369e-02 1.94880564e-03 3.14684039e-05
1.04894288e-01 3.40029150e-02 7.49290809e-02 9.44607367e-04
1.18854735e-02 6.46591525e-06 3.61342798e-04 4.28394735e-04
1.06061827e-06 3.35966423e-03 1.89394839e-02 2.53333170e-02]
Category counts: [ 0 0 1 93 474 0 7 40 44 526 9 10 623 32 9 432 29 0
68 626 22 53 14 337 0 137 6 0 574 165 360 5 59 0 1 2
0 19 101 122]
Current means: [ nan nan 50.504818 44.14718 67.331436 nan 50.613228
48.2564 48.273582 75.28151 46.933273 49.257927 84.50904 56.100006
47.04862 32.123737 46.956173 nan 50.281166 14.506758 49.75228
56.482956 47.379124 60.79012 nan 53.66179 47.505474 nan
24.138206 43.086468 38.52197 51.724556 46.84524 nan 48.63548
50.74783 nan 46.719814 51.802006 56.51408 ]
Posterior variance: [1.0000000e+01 1.0000000e+01 9.0909088e-01 1.0741139e-02 2.1092596e-03
1.0000000e+01 1.4084508e-01 2.4937658e-02 2.2675738e-02 1.9007794e-03
1.0989010e-01 9.9009894e-02 1.6048789e-03 3.1152649e-02 1.0989010e-01
2.3142791e-03 3.4364261e-02 1.0000000e+01 1.4684288e-02 1.5971890e-03
4.5248870e-02 1.8832393e-02 7.0921987e-02 2.9664787e-03 1.0000000e+01
7.2939456e-03 1.6393442e-01 1.0000000e+01 1.7418569e-03 6.0569351e-03
2.7770062e-03 1.9607843e-01 1.6920473e-02 1.0000000e+01 9.0909088e-01
4.7619051e-01 1.0000000e+01 5.2356020e-02 9.8911971e-03 8.1900079e-03]
Posterior means: [ nan nan 50.458927 44.153465 67.32778 nan 50.604588
48.26075 48.2775 75.27671 46.966976 49.26527 84.50351 56.081
47.081055 32.127872 46.966633 nan 50.280754 14.512426 49.7534
56.47075 47.397713 60.786915 nan 53.65912 47.546364 nan
24.14271 43.090656 38.525154 51.690742 46.850582 nan 48.75953
50.712215 nan 46.73699 51.800224 56.508743]
Resampled means: [ nan nan 49.866882 44.157433 67.33187 nan 50.77381
48.251736 48.28439 75.275475 46.77693 49.27621 84.504265 56.040867
46.9859 32.125362 46.945946 nan 50.286263 14.509905 49.760643
56.48236 47.3632 60.786747 nan 53.654045 47.595894 nan
24.14558 43.096577 38.53053 51.786613 46.882935 nan 48.60341
50.012817 nan 46.750515 51.78942 56.488697]
Chosen means: [52.962532 51.18682 49.866882 44.157433 67.33187 50.01998 50.77381
48.251736 48.28439 75.275475 46.77693 49.27621 84.504265 56.040867
46.9859 32.125362 46.945946 50.835552 50.286263 14.509905 49.760643
56.48236 47.3632 60.786747 54.24615 53.654045 47.595894 49.767803
24.14558 43.096577 38.53053 51.786613 46.882935 48.25525 48.60341
50.012817 51.840244 46.750515 51.78942 56.488697]
Category counts: [ 0 1 1 94 477 0 3 40 40 508 12 11 619 37 4 475 37 0
80 625 22 53 6 319 1 145 11 0 523 176 354 3 53 0 2 1
0 24 92 151]
New alpha: [3.125000e-01 1.312500e+00 1.312500e+00 9.431250e+01 4.773125e+02
3.125000e-01 3.312500e+00 4.031250e+01 4.031250e+01 5.083125e+02
1.231250e+01 1.131250e+01 6.193125e+02 3.731250e+01 4.312500e+00
4.753125e+02 3.731250e+01 3.125000e-01 8.031250e+01 6.253125e+02
2.231250e+01 5.331250e+01 6.312500e+00 3.193125e+02 1.312500e+00
1.453125e+02 1.131250e+01 3.125000e-01 5.233125e+02 1.763125e+02
3.543125e+02 3.312500e+00 5.331250e+01 3.125000e-01 2.312500e+00
1.312500e+00 3.125000e-01 2.431250e+01 9.231250e+01 1.513125e+02]
New probs: [2.7061264e-14 1.0102497e-04 3.9649368e-04 1.9037515e-02 9.6542321e-02
6.4238287e-05 5.7832740e-04 7.9233535e-03 6.5184650e-03 1.0014368e-01
2.8404510e-03 1.8451994e-03 1.2255353e-01 6.7550614e-03 7.6169049e-04
9.2134774e-02 5.6979368e-03 4.5110076e-07 1.6616356e-02 1.2751143e-01
4.6129744e-03 7.5059030e-03 1.4483384e-03 6.9339916e-02 1.8508681e-04
2.4075419e-02 2.8154235e-03 5.7850029e-06 1.0112865e-01 3.3934318e-02
7.8736424e-02 6.6373323e-04 1.2288015e-02 2.0643082e-04 2.2716208e-04
6.1714230e-04 8.2841332e-05 3.6348635e-03 1.9542765e-02 3.0926380e-02]
Category counts: [ 0 1 1 94 477 0 3 40 40 508 12 11 619 37 4 475 37 0
80 625 22 53 6 319 1 145 11 0 523 176 354 3 53 0 2 1
0 24 92 151]
Current means: [ nan 50.665203 50.83135 44.195225 67.61912 nan
51.42172 48.01704 48.282818 75.512596 46.831165 49.0099
84.53707 55.985535 47.075928 31.647778 46.945618 nan
50.342705 14.4979315 49.876915 56.937115 47.37717 61.258186
52.22962 53.597412 47.45203 nan 23.705576 42.810177
38.3593 51.933884 46.788887 nan 47.867302 50.76721
nan 46.807358 51.822357 56.94306 ]
Posterior variance: [1.0000000e+01 9.0909088e-01 9.0909088e-01 1.0626993e-02 2.0959966e-03
1.0000000e+01 3.2258067e-01 2.4937658e-02 2.4937658e-02 1.9681165e-03
8.2644626e-02 9.0090089e-02 1.6152480e-03 2.6954180e-02 2.4390244e-01
2.1048200e-03 2.6954180e-02 1.0000000e+01 1.2484395e-02 1.5997441e-03
4.5248870e-02 1.8832393e-02 1.6393442e-01 3.1338139e-03 9.0909088e-01
6.8917987e-03 9.0090089e-02 1.0000000e+01 1.9116804e-03 5.6785913e-03
2.8240609e-03 3.2258067e-01 1.8832393e-02 1.0000000e+01 4.7619051e-01
9.0909088e-01 1.0000000e+01 4.1493773e-02 1.0857764e-02 6.6181333e-03]
Posterior means: [ nan 50.604733 50.755775 44.201393 67.61543 nan 51.37586
48.021984 48.2871 75.507576 46.857357 49.018818 84.531494 55.9694
47.147247 31.65164 46.95385 nan 50.34228 14.503611 49.877472
56.924053 47.420166 61.254658 52.02693 53.59493 47.474987 nan
23.710602 42.81426 38.362587 51.871502 46.794937 nan 47.968857
50.697464 nan 46.820602 51.820374 56.93847 ]
Resampled means: [ nan 50.886185 50.256966 44.19914 67.61851 nan 51.488247
48.002243 48.29867 75.507126 46.863827 49.04138 84.53141 55.98984
47.232273 31.651987 46.973427 nan 50.346798 14.501753 49.893887
56.937798 47.26357 61.257084 51.98892 53.59759 47.331974 nan
23.71408 42.8062 38.36023 52.4707 46.77241 nan 48.62526
51.379192 nan 46.854176 51.816055 56.93504 ]
Chosen means: [52.962532 50.886185 50.256966 44.19914 67.61851 50.01998 51.488247
48.002243 48.29867 75.507126 46.863827 49.04138 84.53141 55.98984
47.232273 31.651987 46.973427 50.835552 50.346798 14.501753 49.893887
56.937798 47.26357 61.257084 51.98892 53.59759 47.331974 49.767803
23.71408 42.8062 38.36023 52.4707 46.77241 48.25525 48.62526
51.379192 51.840244 46.854176 51.816055 56.93504 ]
Category counts: [ 0 0 0 109 449 1 1 37 31 513 14 11 612 38 4 483 35 0
74 618 30 41 8 329 0 144 15 0 508 168 359 2 59 0 1 4
0 21 110 171]
New alpha: [3.125000e-01 3.125000e-01 3.125000e-01 1.093125e+02 4.493125e+02
1.312500e+00 1.312500e+00 3.731250e+01 3.131250e+01 5.133125e+02
1.431250e+01 1.131250e+01 6.123125e+02 3.831250e+01 4.312500e+00
4.833125e+02 3.531250e+01 3.125000e-01 7.431250e+01 6.183125e+02
3.031250e+01 4.131250e+01 8.312500e+00 3.293125e+02 3.125000e-01
1.443125e+02 1.531250e+01 3.125000e-01 5.083125e+02 1.683125e+02
3.593125e+02 2.312500e+00 5.931250e+01 3.125000e-01 1.312500e+00
4.312500e+00 3.125000e-01 2.131250e+01 1.103125e+02 1.713125e+02]
New probs: [7.2184761e-05 6.7532572e-05 1.7559207e-05 2.3426261e-02 9.4435208e-02
4.6346649e-05 1.9206892e-04 7.3551997e-03 4.9451706e-03 1.0019397e-01
2.8814392e-03 3.2654097e-03 1.2278604e-01 7.6588392e-03 9.4267342e-04
9.8728180e-02 6.6081369e-03 3.1564072e-05 1.0927072e-02 1.3219835e-01
6.9093662e-03 6.1615729e-03 2.4984267e-03 5.8352184e-02 1.4219140e-06
2.8873231e-02 1.6093648e-03 3.0302361e-04 1.0118557e-01 3.3103149e-02
7.6049037e-02 2.4088955e-04 1.2240850e-02 4.2394118e-04 5.7477353e-04
9.9227740e-04 9.6257224e-07 4.8153475e-03 1.8276764e-02 3.0608665e-02]
Category counts: [ 0 0 0 109 449 1 1 37 31 513 14 11 612 38 4 483 35 0
74 618 30 41 8 329 0 144 15 0 508 168 359 2 59 0 1 4
0 21 110 171]
Current means: [ nan nan nan 44.128983 67.84847 50.338783 50.64495
48.13268 48.32005 75.59316 46.94513 48.704247 84.587715 55.583908
47.467407 31.394184 46.812496 nan 50.29608 14.447742 49.92968
57.55511 47.578293 61.706543 nan 53.847504 46.960564 nan
23.44825 42.56272 38.164856 53.207382 46.923294 nan 46.041267
51.181522 nan 46.867603 51.76486 57.19459 ]
Posterior variance: [1.0000000e+01 1.0000000e+01 1.0000000e+01 9.1659026e-03 2.2266756e-03
9.0909088e-01 9.0909088e-01 2.6954180e-02 3.2154340e-02 1.9489379e-03
7.0921987e-02 9.0090089e-02 1.6337200e-03 2.6246721e-02 2.4390244e-01
2.0699648e-03 2.8490029e-02 1.0000000e+01 1.3495277e-02 1.6178612e-03
3.3222590e-02 2.4330901e-02 1.2345678e-01 3.0385901e-03 1.0000000e+01
6.9396249e-03 6.6225164e-02 1.0000000e+01 1.9681165e-03 5.9488397e-03
2.7847395e-03 4.7619051e-01 1.6920473e-02 1.0000000e+01 9.0909088e-01
2.4390244e-01 1.0000000e+01 4.7393363e-02 9.0826526e-03 5.8445353e-03]
Posterior means: [ nan nan nan 44.134365 67.8445 50.307987 50.586323
48.13771 48.32545 75.58817 46.966797 48.71592 84.58207 55.569252
47.52918 31.398035 46.821575 nan 50.295685 14.453493 49.929913
57.53673 47.60819 61.702988 nan 53.844833 46.980694 nan
23.453474 42.567142 38.16815 53.05465 46.9285 nan 46.401154
51.152706 nan 46.882446 51.763252 57.19039 ]
Resampled means: [ nan nan nan 44.130833 67.84028 50.533836 51.756172
48.129883 48.31738 75.587 46.836655 48.73281 84.58114 55.554356
47.472233 31.399939 46.85263 nan 50.273735 14.455692 49.948486
57.52231 47.7081 61.69536 nan 53.85037 47.063522 nan
23.452818 42.57208 38.168026 52.75849 46.902187 nan 44.909096
51.156498 nan 46.833946 51.761314 57.19112 ]
Chosen means: [52.962532 50.886185 50.256966 44.130833 67.84028 50.533836 51.756172
48.129883 48.31738 75.587 46.836655 48.73281 84.58114 55.554356
47.472233 31.399939 46.85263 50.835552 50.273735 14.455692 49.948486
57.52231 47.7081 61.69536 51.98892 53.85037 47.063522 49.767803
23.452818 42.57208 38.168026 52.75849 46.902187 48.25525 44.909096
51.156498 51.840244 46.833946 51.761314 57.19112 ]
Category counts: [ 0 3 0 117 417 1 1 47 22 513 14 20 609 42 7 467 34 0
64 615 41 37 13 336 0 151 8 3 509 187 348 3 48 0 1 1
0 32 101 188]
New alpha: [3.125000e-01 3.312500e+00 3.125000e-01 1.173125e+02 4.173125e+02
1.312500e+00 1.312500e+00 4.731250e+01 2.231250e+01 5.133125e+02
1.431250e+01 2.031250e+01 6.093125e+02 4.231250e+01 7.312500e+00
4.673125e+02 3.431250e+01 3.125000e-01 6.431250e+01 6.153125e+02
4.131250e+01 3.731250e+01 1.331250e+01 3.363125e+02 3.125000e-01
1.513125e+02 8.312500e+00 3.312500e+00 5.093125e+02 1.873125e+02
3.483125e+02 3.312500e+00 4.831250e+01 3.125000e-01 1.312500e+00
1.312500e+00 3.125000e-01 3.231250e+01 1.013125e+02 1.883125e+02]
New probs: [2.88106935e-06 1.52018154e-04 3.44513028e-05 2.70696767e-02
8.32427442e-02 7.74487780e-05 4.45622136e-05 8.31453130e-03
3.50612449e-03 9.77649987e-02 2.90313596e-03 3.17832804e-03
1.27452344e-01 8.19270965e-03 1.81918754e-03 8.89649093e-02
7.14152632e-03 2.70216347e-04 1.12701990e-02 1.23697855e-01
8.37881118e-03 9.91108548e-03 4.17949120e-03 6.56166300e-02
7.39883180e-05 3.13556194e-02 1.14286214e-03 4.55620844e-04
9.03932303e-02 4.31565754e-02 7.43773878e-02 1.31575123e-03
1.08288871e-02 2.67955911e-05 9.20890816e-05 3.06321715e-04
1.44504302e-04 5.33681270e-03 2.17073299e-02 3.61000784e-02]
Category counts: [ 0 3 0 117 417 1 1 47 22 513 14 20 609 42 7 467 34 0
64 615 41 37 13 336 0 151 8 3 509 187 348 3 48 0 1 1
0 32 101 188]
Current means: [ nan 51.920284 nan 44.37426 68.16612 50.50239 52.383488
48.32953 48.37885 75.642395 46.883125 49.194 84.60794 55.77404
47.728447 31.25182 46.745476 nan 50.302452 14.425735 50.08545
57.962975 47.39484 62.216373 nan 53.82524 47.233322 49.21819
23.406683 42.045944 37.78721 52.925518 46.632603 nan 44.681355
50.984196 nan 46.72364 52.019375 57.549618]
Posterior variance: [1.00000000e+01 3.22580665e-01 1.00000000e+01 8.53971019e-03
2.39750650e-03 9.09090877e-01 9.09090877e-01 2.12314241e-02
4.52488698e-02 1.94893789e-03 7.09219873e-02 4.97512445e-02
1.64176663e-03 2.37529706e-02 1.40845075e-01 2.14086915e-03
2.93255150e-02 1.00000000e+01 1.56006245e-02 1.62575196e-03
2.43309010e-02 2.69541796e-02 7.63358772e-02 2.97530484e-03
1.00000000e+01 6.61813328e-03 1.23456784e-01 3.22580665e-01
1.96425058e-03 5.34473546e-03 2.87273759e-03 3.22580665e-01
2.07900219e-02 1.00000000e+01 9.09090877e-01 9.09090877e-01
1.00000000e+01 3.11526489e-02 9.89119709e-03 5.31632081e-03]
Posterior means: [ nan 51.85834 nan 44.379063 68.161766 50.456722 52.16681
48.333076 48.386185 75.6374 46.90523 49.19801 84.60226 55.760323
47.760437 31.255835 46.755024 nan 50.30198 14.431519 50.085243
57.94151 47.414726 62.212738 nan 53.822712 47.26748 49.243412
23.411907 42.050198 37.790718 52.83115 46.639606 nan 45.16487
50.894726 nan 46.73385 52.01738 57.5456 ]
Resampled means: [ nan 51.324306 nan 44.376503 68.15907 50.699078 52.886883
48.36277 48.34537 75.63562 46.873127 49.13327 84.60377 55.77431
47.84722 31.257137 46.734478 nan 50.270996 14.431129 50.080894
57.947437 47.54765 62.208347 nan 53.827984 47.44052 49.51603
23.40898 42.05283 37.797546 52.545055 46.646217 nan 45.787285
51.148026 nan 46.754395 52.019157 57.53841 ]
Chosen means: [52.962532 51.324306 50.256966 44.376503 68.15907 50.699078 52.886883
48.36277 48.34537 75.63562 46.873127 49.13327 84.60377 55.77431
47.84722 31.257137 46.734478 50.835552 50.270996 14.431129 50.080894
57.947437 47.54765 62.208347 51.98892 53.827984 47.44052 49.51603
23.40898 42.05283 37.797546 52.545055 46.646217 48.25525 45.787285
51.148026 51.840244 46.754395 52.019157 57.53841 ]
Category counts: [ 0 1 0 123 383 1 1 37 17 514 11 21 606 50 12 441 30 1
62 617 31 52 23 366 0 163 9 2 507 217 337 4 70 0 1 2
0 20 104 164]
New alpha: [3.125000e-01 1.312500e+00 3.125000e-01 1.233125e+02 3.833125e+02
1.312500e+00 1.312500e+00 3.731250e+01 1.731250e+01 5.143125e+02
1.131250e+01 2.131250e+01 6.063125e+02 5.031250e+01 1.231250e+01
4.413125e+02 3.031250e+01 1.312500e+00 6.231250e+01 6.173125e+02
3.131250e+01 5.231250e+01 2.331250e+01 3.663125e+02 3.125000e-01
1.633125e+02 9.312500e+00 2.312500e+00 5.073125e+02 2.173125e+02
3.373125e+02 4.312500e+00 7.031250e+01 3.125000e-01 1.312500e+00
2.312500e+00 3.125000e-01 2.031250e+01 1.043125e+02 1.643125e+02]
New probs: [2.18680580e-04 3.65563668e-04 2.08969068e-05 2.63083763e-02
7.73552209e-02 1.00433317e-04 2.29720798e-04 9.27761756e-03
2.02215137e-03 1.06505141e-01 2.30623758e-03 5.33590931e-03
1.16724677e-01 1.09118475e-02 1.71987235e-03 8.95362496e-02
9.07636341e-03 2.61166628e-04 1.16285020e-02 1.17904104e-01
6.38938183e-03 9.25990753e-03 4.38194489e-03 6.95883408e-02
1.57432296e-05 3.36295143e-02 1.87595235e-03 1.30905572e-03
1.01365454e-01 4.36724350e-02 6.92752078e-02 9.39358317e-04
1.36687178e-02 1.80969888e-04 2.94700236e-04 4.59343428e-05
2.19483276e-08 4.03690245e-03 1.99624207e-02 3.22987773e-02]
Category counts: [ 0 1 0 123 383 1 1 37 17 514 11 21 606 50 12 441 30 1
62 617 31 52 23 366 0 163 9 2 507 217 337 4 70 0 1 2
0 20 104 164]
Current means: [ nan 50.559174 nan 44.32967 68.51212 50.493614 52.27214
48.45697 48.423725 75.68019 47.13719 48.94205 84.63175 55.883125
47.685627 31.048399 46.83685 52.178345 50.371357 14.441557 50.186546
58.02097 47.666798 62.513256 nan 53.92306 47.308746 48.37358
23.42206 41.686703 37.25731 52.565727 46.75865 nan 46.429462
51.548576 nan 46.35463 51.840897 57.686813]
Posterior variance: [1.0000000e+01 9.0909088e-01 1.0000000e+01 8.1234770e-03 2.6102844e-03
9.0909088e-01 9.0909088e-01 2.6954180e-02 5.8479533e-02 1.9451469e-03
9.0090089e-02 4.7393363e-02 1.6498928e-03 1.9960081e-02 8.2644626e-02
2.2670596e-03 3.3222590e-02 9.0909088e-01 1.6103059e-02 1.6204830e-03
3.2154340e-02 1.9193858e-02 4.3290041e-02 2.7314941e-03 1.0000000e+01
6.1312076e-03 1.0989010e-01 4.7619051e-01 1.9719976e-03 4.6061720e-03
2.9664787e-03 2.4390244e-01 1.4265335e-02 1.0000000e+01 9.0909088e-01
4.7619051e-01 1.0000000e+01 4.9751244e-02 9.6061481e-03 6.0938448e-03]
Posterior means: [ nan 50.508343 nan 44.33428 68.50729 50.448742 52.065586
48.46113 48.43294 75.6752 47.162983 48.947063 84.62604 55.87138
47.704758 31.052694 46.84736 51.980316 50.370758 14.44732 50.185944
58.005573 47.6769 62.509834 nan 53.92066 47.33832 48.451027
23.427301 41.69053 37.261086 52.50315 46.763275 nan 46.75406
51.474834 nan 46.372765 51.83913 57.68213 ]
Resampled means: [ nan 51.47635 nan 44.33099 68.50866 51.120132 51.656677
48.45733 48.4433 75.67171 47.227142 48.89718 84.627174 55.865555
47.713192 31.055859 46.850872 52.59696 50.348305 14.447564 50.186478
58.01448 47.63015 62.510685 nan 53.919315 47.04642 48.540627
23.42967 41.686077 37.262573 52.18765 46.74556 nan 45.1639
51.90899 nan 46.314117 51.831303 57.676975]
Chosen means: [52.962532 51.47635 50.256966 44.33099 68.50866 51.120132 51.656677
48.45733 48.4433 75.67171 47.227142 48.89718 84.627174 55.865555
47.713192 31.055859 46.850872 52.59696 50.348305 14.447564 50.186478
58.01448 47.63015 62.510685 51.98892 53.919315 47.04642 48.540627
23.42967 41.686077 37.262573 52.18765 46.74556 48.25525 45.1639
51.90899 51.840244 46.314117 51.831303 57.676975]
Category counts: [ 1 1 0 144 386 1 0 43 7 508 10 33 604 63 6 406 53 2
60 612 26 53 19 359 0 159 7 2 509 226 350 7 58 0 3 0
0 18 102 162]
New alpha: [1.312500e+00 1.312500e+00 3.125000e-01 1.443125e+02 3.863125e+02
1.312500e+00 3.125000e-01 4.331250e+01 7.312500e+00 5.083125e+02
1.031250e+01 3.331250e+01 6.043125e+02 6.331250e+01 6.312500e+00
4.063125e+02 5.331250e+01 2.312500e+00 6.031250e+01 6.123125e+02
2.631250e+01 5.331250e+01 1.931250e+01 3.593125e+02 3.125000e-01
1.593125e+02 7.312500e+00 2.312500e+00 5.093125e+02 2.263125e+02
3.503125e+02 7.312500e+00 5.831250e+01 3.125000e-01 3.312500e+00
3.125000e-01 3.125000e-01 1.831250e+01 1.023125e+02 1.623125e+02]
New probs: [6.13181182e-05 3.09520168e-04 3.07778478e-08 2.60262862e-02
7.55651295e-02 2.23085342e-04 5.94459243e-06 8.09589867e-03
9.73823131e-04 1.03210062e-01 1.84325071e-03 5.96961565e-03
1.14085093e-01 1.17808394e-02 1.79958367e-03 7.88100138e-02
1.09924078e-02 7.25794671e-05 1.30708665e-02 1.32914364e-01
4.61537251e-03 1.06584392e-02 5.42533398e-03 7.25701004e-02
3.11878030e-05 3.34164836e-02 1.45495380e-03 6.37155608e-04
1.12550065e-01 4.47362661e-02 6.57488331e-02 1.64330937e-03
9.74102318e-03 4.65584963e-05 1.07976503e-03 1.27418476e-04
1.62076700e-04 4.55974042e-03 1.91918239e-02 2.57946551e-02]
Category counts: [ 1 1 0 144 386 1 0 43 7 508 10 33 604 63 6 406 53 2
60 612 26 53 19 359 0 159 7 2 509 226 350 7 58 0 3 0
0 18 102 162]
Current means: [54.3871 50.55954 nan 44.317703 68.630325 50.30497 nan
48.620304 48.929607 75.75384 47.935772 49.126724 84.64553 55.98511
48.347797 30.701487 46.603413 52.59606 50.69306 14.405601 50.3466
58.208164 47.945 62.625423 nan 53.999004 46.72098 48.532288
23.357365 41.40605 36.737125 52.45171 46.86581 nan 46.04463
nan nan 46.474888 51.66863 57.87637 ]
Posterior variance: [9.0909088e-01 9.0909088e-01 1.0000000e+01 6.9396249e-03 2.5900025e-03
9.0909088e-01 1.0000000e+01 2.3201857e-02 1.4084508e-01 1.9681165e-03
9.9009894e-02 3.0211482e-02 1.6553551e-03 1.5847862e-02 1.6393442e-01
2.4624476e-03 1.8832393e-02 4.7619051e-01 1.6638935e-02 1.6337200e-03
3.8314175e-02 1.8832393e-02 5.2356020e-02 2.7847395e-03 1.0000000e+01
6.2853550e-03 1.4084508e-01 4.7619051e-01 1.9642506e-03 4.4228216e-03
2.8563268e-03 1.4084508e-01 1.7211704e-02 1.0000000e+01 3.2258067e-01
1.0000000e+01 1.0000000e+01 5.5248618e-02 9.7943190e-03 6.1690314e-03]
Posterior means: [53.988277 50.508675 nan 44.321648 68.625496 50.27725 nan
48.62351 48.94468 75.74877 47.956207 49.129364 84.6398 55.975628
48.37488 30.706238 46.60981 52.47244 50.69191 14.411416 50.34527
58.192707 47.95576 62.621906 nan 53.99649 46.767162 48.602177
23.362598 41.40985 36.740913 52.417175 46.871204 nan 46.172222
nan nan 46.494366 51.666996 57.87151 ]
Resampled means: [54.785324 50.362576 nan 44.331 68.62418 49.09445 nan
48.60738 48.81573 75.7471 47.928257 49.175262 84.6419 55.950485
48.471313 30.706743 46.609196 53.070488 50.673042 14.408944 50.355675
58.185616 47.99141 62.622765 nan 53.995655 46.570087 48.61542
23.363451 41.41095 36.748146 52.456608 46.838444 nan 45.713333
nan nan 46.55952 51.664837 57.874477]
Chosen means: [54.785324 50.362576 50.256966 44.331 68.62418 49.09445 51.656677
48.60738 48.81573 75.7471 47.928257 49.175262 84.6419 55.950485
48.471313 30.706743 46.609196 53.070488 50.673042 14.408944 50.355675
58.185616 47.99141 62.622765 51.98892 53.995655 46.570087 48.61542
23.363451 41.41095 36.748146 52.456608 46.838444 48.25525 45.713333
51.90899 51.840244 46.55952 51.664837 57.874477]
Category counts: [ 0 0 0 144 387 2 0 42 0 511 4 27 600 74 15 388 66 1
55 612 28 70 32 347 0 169 4 3 504 239 360 9 51 0 6 1
0 16 87 146]
New alpha: [3.125000e-01 3.125000e-01 3.125000e-01 1.443125e+02 3.873125e+02
2.312500e+00 3.125000e-01 4.231250e+01 3.125000e-01 5.113125e+02
4.312500e+00 2.731250e+01 6.003125e+02 7.431250e+01 1.531250e+01
3.883125e+02 6.631250e+01 1.312500e+00 5.531250e+01 6.123125e+02
2.831250e+01 7.031250e+01 3.231250e+01 3.473125e+02 3.125000e-01
1.693125e+02 4.312500e+00 3.312500e+00 5.043125e+02 2.393125e+02
3.603125e+02 9.312500e+00 5.131250e+01 3.125000e-01 6.312500e+00
1.312500e+00 3.125000e-01 1.631250e+01 8.731250e+01 1.463125e+02]
New probs: [5.5022940e-07 3.0267096e-04 6.3524590e-07 2.5555480e-02 7.6980419e-02
3.8213024e-04 2.2580285e-04 9.7997943e-03 1.5557227e-04 1.0730188e-01
1.2567432e-03 6.2593408e-03 1.2909627e-01 1.3676280e-02 2.3612233e-03
7.6512471e-02 1.2365999e-02 4.3235955e-04 1.1849331e-02 1.1437607e-01
5.3246575e-03 1.4966677e-02 7.7714068e-03 6.8627082e-02 2.5861489e-05
3.4210093e-02 1.0923138e-03 9.7642234e-04 9.8569468e-02 4.6043143e-02
7.4139550e-02 2.2035085e-03 8.7487129e-03 5.3070381e-04 6.9546833e-04
2.7074464e-04 3.6979185e-05 2.6738311e-03 1.5866427e-02 2.8336383e-02]
Category counts: [ 0 0 0 144 387 2 0 42 0 511 4 27 600 74 15 388 66 1
55 612 28 70 32 347 0 169 4 3 504 239 360 9 51 0 6 1
0 16 87 146]
Current means: [ nan nan nan 44.282413 68.63991 48.861847 nan
48.586407 nan 75.79298 49.593864 49.029007 84.676704 56.07272
48.902897 30.439249 46.74512 52.309032 50.762714 14.405591 50.465843
58.44057 47.96418 62.70321 nan 53.952553 47.039585 48.471508
23.323893 41.30079 36.492466 52.45295 46.961685 nan 46.240482
50.837963 nan 46.528984 51.62077 58.051525]
Posterior variance: [1.0000000e+01 1.0000000e+01 1.0000000e+01 6.9396249e-03 2.5833119e-03
4.7619051e-01 1.0000000e+01 2.3752971e-02 1.0000000e+01 1.9565641e-03
2.4390244e-01 3.6900368e-02 1.6663891e-03 1.3495277e-02 6.6225164e-02
2.5766555e-03 1.5128593e-02 9.0909088e-01 1.8148821e-02 1.6337200e-03
3.5587188e-02 1.4265335e-02 3.1152649e-02 2.8810140e-03 1.0000000e+01
5.9136604e-03 2.4390244e-01 3.2258067e-01 1.9837334e-03 4.1823504e-03
2.7770062e-03 1.0989010e-01 1.9569471e-02 1.0000000e+01 1.6393442e-01
9.0909088e-01 1.0000000e+01 6.2111799e-02 1.1481057e-02 6.8446267e-03]
Posterior means: [ nan nan nan 44.28638 68.63509 48.916042
nan 48.589764 nan 75.787926 49.60377 49.032593
84.67092 56.064526 48.910164 30.444288 46.750042 52.099125
50.76133 14.4114065 50.464188 58.428528 47.970524 62.699547
nan 53.950214 47.111794 48.520817 23.329184 41.30443
36.496216 52.425995 46.967632 nan 46.302113 50.761787
nan 46.550545 51.618904 58.046013 ]
Resampled means: [ nan nan nan 44.293953 68.63869 48.450962 nan
48.620197 nan 75.78464 49.74146 49.001488 84.67144 56.06142
48.852425 30.443342 46.73002 53.002316 50.779625 14.410494 50.44956
58.400265 47.98959 62.696518 nan 53.962646 47.183113 48.478283
23.329332 41.30088 36.49556 52.604538 46.95366 nan 46.496227
50.840443 nan 46.470547 51.62973 58.052776]
Chosen means: [54.785324 50.362576 50.256966 44.293953 68.63869 48.450962 51.656677
48.620197 48.81573 75.78464 49.74146 49.001488 84.67144 56.06142
48.852425 30.443342 46.73002 53.002316 50.779625 14.410494 50.44956
58.400265 47.98959 62.696518 51.98892 53.962646 47.183113 48.478283
23.329332 41.30088 36.49556 52.604538 46.95366 48.25525 46.496227
50.840443 51.840244 46.470547 51.62973 58.052776]
Category counts: [ 0 1 0 140 400 0 0 49 1 506 10 28 593 90 9 388 64 2
70 608 28 77 34 328 0 162 8 5 503 245 357 12 44 1 6 4
0 14 71 142]
New alpha: [3.125000e-01 1.312500e+00 3.125000e-01 1.403125e+02 4.003125e+02
3.125000e-01 3.125000e-01 4.931250e+01 1.312500e+00 5.063125e+02
1.031250e+01 2.831250e+01 5.933125e+02 9.031250e+01 9.312500e+00
3.883125e+02 6.431250e+01 2.312500e+00 7.031250e+01 6.083125e+02
2.831250e+01 7.731250e+01 3.431250e+01 3.283125e+02 3.125000e-01
1.623125e+02 8.312500e+00 5.312500e+00 5.033125e+02 2.453125e+02
3.573125e+02 1.231250e+01 4.431250e+01 1.312500e+00 6.312500e+00
4.312500e+00 3.125000e-01 1.431250e+01 7.131250e+01 1.423125e+02]
New probs: [8.65303155e-05 1.07864886e-04 1.22180239e-07 2.94431131e-02
8.23498666e-02 2.63679802e-04 3.16001979e-06 8.67303368e-03
1.01903963e-04 1.08072788e-01 1.54459744e-03 4.93239472e-03
1.20051794e-01 1.65823121e-02 1.99768948e-03 7.67828226e-02
1.30979568e-02 2.48043216e-04 1.19200209e-02 1.22810282e-01
5.37806470e-03 1.58272851e-02 6.96109375e-03 5.89160807e-02
1.11304384e-04 3.57171856e-02 1.55356701e-03 1.25211233e-03
9.67356339e-02 4.75406460e-02 7.14384615e-02 2.03004666e-03
1.02129392e-02 9.92069719e-04 1.98390405e-03 5.62202302e-04
1.15492078e-07 2.21090531e-03 1.44404713e-02 2.70660110e-02]
Category counts: [ 0 1 0 140 400 0 0 49 1 506 10 28 593 90 9 388 64 2
70 608 28 77 34 328 0 162 8 5 503 245 357 12 44 1 6 4
0 14 71 142]
Current means: [ nan 48.677013 nan 44.258507 68.74148 nan nan
48.541954 48.374683 75.93782 49.58125 48.93934 84.729935 56.051
48.688225 30.354279 46.465496 52.450592 50.728645 14.375331 50.836857
58.902634 47.92643 62.81496 nan 53.957386 46.950874 47.955303
23.25353 41.18593 36.40016 52.650166 47.088028 48.392933 46.459087
50.88317 nan 46.087627 51.77462 58.270557]
Posterior variance: [1.00000000e+01 9.09090877e-01 1.00000000e+01 7.13775866e-03
2.49937503e-03 1.00000000e+01 1.00000000e+01 2.03665998e-02
9.09090877e-01 1.97589397e-03 9.90098938e-02 3.55871879e-02
1.68605638e-03 1.10987797e-02 1.09890103e-01 2.57665548e-03
1.56006245e-02 4.76190507e-01 1.42653352e-02 1.64446642e-03
3.55871879e-02 1.29701691e-02 2.93255150e-02 3.04785115e-03
1.00000000e+01 6.16903137e-03 1.23456784e-01 1.96078435e-01
1.98767637e-03 4.07996727e-03 2.80033611e-03 8.26446265e-02
2.26757377e-02 9.09090877e-01 1.63934425e-01 2.43902445e-01
1.00000000e+01 7.09219873e-02 1.40646975e-02 7.03729736e-03]
Posterior means: [ nan 48.797287 nan 44.262608 68.73679 nan nan
48.54492 48.52244 75.93269 49.585392 48.943115 84.72408 56.04428
48.70264 30.359339 46.47101 52.333897 50.727604 14.381188 50.83388
58.891083 47.932514 62.811054 nan 53.954945 46.988518 47.995396
23.258848 41.189526 36.40397 52.628265 47.094635 48.539032 46.51713
50.861633 nan 46.115376 51.772125 58.26474 ]
Resampled means: [ nan 48.72506 nan 44.268845 68.73629 nan nan
48.53182 47.863136 75.929985 49.62591 48.951447 84.72761 56.028175
48.478615 30.357594 46.456688 51.997906 50.72084 14.38114 50.798203
58.890335 47.908867 62.812965 nan 53.94899 46.9622 47.79916
23.260815 41.192127 36.40196 52.73834 47.065575 48.437866 46.705574
50.72838 nan 46.1161 51.78049 58.25899 ]
Chosen means: [54.785324 48.72506 50.256966 44.268845 68.73629 48.450962 51.656677
48.53182 47.863136 75.929985 49.62591 48.951447 84.72761 56.028175
48.478615 30.357594 46.456688 51.997906 50.72084 14.38114 50.798203
58.890335 47.908867 62.812965 51.98892 53.94899 46.9622 47.79916
23.260815 41.192127 36.40196 52.73834 47.065575 48.437866 46.705574
50.72838 51.840244 46.1161 51.78049 58.25899 ]
Category counts: [ 0 0 0 162 406 1 0 48 0 510 5 27 580 103 7 392 60 1
75 606 22 94 25 306 1 157 5 6 499 250 345 14 55 4 12 5
0 5 75 137]
New alpha: [3.125000e-01 3.125000e-01 3.125000e-01 1.623125e+02 4.063125e+02
1.312500e+00 3.125000e-01 4.831250e+01 3.125000e-01 5.103125e+02
5.312500e+00 2.731250e+01 5.803125e+02 1.033125e+02 7.312500e+00
3.923125e+02 6.031250e+01 1.312500e+00 7.531250e+01 6.063125e+02
2.231250e+01 9.431250e+01 2.531250e+01 3.063125e+02 1.312500e+00
1.573125e+02 5.312500e+00 6.312500e+00 4.993125e+02 2.503125e+02
3.453125e+02 1.431250e+01 5.531250e+01 4.312500e+00 1.231250e+01
5.312500e+00 3.125000e-01 5.312500e+00 7.531250e+01 1.373125e+02]
New probs: [4.9959187e-08 6.5020358e-09 4.1206917e-08 3.4166224e-02 8.0338001e-02
4.9619435e-04 3.2374068e-05 9.3826829e-03 5.1125240e-05 1.1110215e-01
1.1903753e-03 4.9133571e-03 1.1225811e-01 1.6696367e-02 1.7124411e-03
8.4942862e-02 1.3559049e-02 7.4127027e-05 1.3177190e-02 1.1769284e-01
3.4338483e-03 2.2185523e-02 5.3121876e-03 5.9541505e-02 1.5263635e-05
3.3372790e-02 7.3269889e-04 1.5462791e-03 9.6982867e-02 5.0306659e-02
6.5639980e-02 2.5012041e-03 1.0402044e-02 1.1461738e-03 1.6975332e-03
9.7154395e-04 1.3496401e-04 9.8722242e-04 1.6126288e-02 2.5177965e-02]
Category counts: [ 0 0 0 162 406 1 0 48 0 510 5 27 580 103 7 392 60 1
75 606 22 94 25 306 1 157 5 6 499 250 345 14 55 4 12 5
0 5 75 137]
Current means: [ nan nan nan 44.26426 68.8368 50.236065 nan
48.477955 nan 76.11437 50.150463 49.20505 84.828384 56.16548
48.921223 30.281796 46.72175 52.76932 50.819344 14.362005 50.695602
59.386803 47.876183 63.042503 54.09299 53.920414 46.611393 48.271236
23.192081 40.98868 36.290417 52.727264 47.082947 48.519367 46.799526
50.996468 nan 45.89495 51.883 58.425987]
Posterior variance: [1.0000000e+01 1.0000000e+01 1.0000000e+01 6.1690314e-03 2.4624476e-03
9.0909088e-01 1.0000000e+01 2.0790022e-02 1.0000000e+01 1.9603998e-03
1.9607843e-01 3.6900368e-02 1.7238408e-03 9.6993214e-03 1.4084508e-01
2.5503698e-03 1.6638935e-02 9.0909088e-01 1.3315580e-02 1.6498928e-03
4.5248870e-02 1.0626993e-02 3.9840639e-02 3.2669061e-03 9.0909088e-01
6.3653723e-03 1.9607843e-01 1.6393442e-01 2.0036064e-03 3.9984006e-03
2.8977108e-03 7.0921987e-02 1.8148821e-02 2.4390244e-01 8.2644626e-02
1.9607843e-01 1.0000000e+01 1.9607843e-01 1.3315580e-02 7.2939456e-03]
Posterior means: [ nan nan nan 44.2678 68.83216 50.214607 nan
48.48112 nan 76.10925 50.147514 49.207985 84.82238 56.159504
48.936413 30.286825 46.727203 52.517567 50.818253 14.367885 50.692455
59.376827 47.884647 63.038242 53.720905 53.917915 46.677837 48.299576
23.197453 40.992287 36.29439 52.707924 47.08824 48.55548 46.825977
50.97693 nan 45.97544 51.880493 58.41984 ]
Resampled means: [ nan nan nan 44.26481 68.83198 49.345673 nan
48.470993 nan 76.10928 49.913254 49.12682 84.82178 56.149105
48.968414 30.286142 46.716167 50.389515 50.818615 14.368418 50.73428
59.379944 47.861496 63.034462 55.074005 53.91642 46.808556 48.34504
23.195333 40.988327 36.29487 52.71478 47.09159 48.283222 46.81593
50.776936 nan 45.950222 51.898148 58.426094]
Chosen means: [54.785324 48.72506 50.256966 44.26481 68.83198 49.345673 51.656677
48.470993 47.863136 76.10928 49.913254 49.12682 84.82178 56.149105
48.968414 30.286142 46.716167 50.389515 50.818615 14.368418 50.73428
59.379944 47.861496 63.034462 55.074005 53.91642 46.808556 48.34504
23.195333 40.988327 36.29487 52.71478 47.09159 48.283222 46.81593
50.776936 51.840244 45.950222 51.898148 58.426094]
Category counts: [ 0 0 0 173 419 2 0 49 0 515 8 18 560 115 14 407 73 0
52 595 19 125 21 278 0 154 3 7 494 260 330 15 41 10 9 5
0 4 101 124]
New alpha: [3.125000e-01 3.125000e-01 3.125000e-01 1.733125e+02 4.193125e+02
2.312500e+00 3.125000e-01 4.931250e+01 3.125000e-01 5.153125e+02
8.312500e+00 1.831250e+01 5.603125e+02 1.153125e+02 1.431250e+01
4.073125e+02 7.331250e+01 3.125000e-01 5.231250e+01 5.953125e+02
1.931250e+01 1.253125e+02 2.131250e+01 2.783125e+02 3.125000e-01
1.543125e+02 3.312500e+00 7.312500e+00 4.943125e+02 2.603125e+02
3.303125e+02 1.531250e+01 4.131250e+01 1.031250e+01 9.312500e+00
5.312500e+00 3.125000e-01 4.312500e+00 1.013125e+02 1.243125e+02]
New probs: [5.74024625e-06 3.15531332e-04 3.57416830e-05 3.27434428e-02
9.07265991e-02 3.86525528e-04 1.71038373e-05 1.20109795e-02
1.76165292e-06 9.75727513e-02 6.40950922e-04 3.28610628e-03
1.02545552e-01 2.14902870e-02 2.21244828e-03 9.25975740e-02
1.68150216e-02 2.54406160e-07 9.86529328e-03 1.24923728e-01
2.85546109e-03 2.24736109e-02 3.63580673e-03 5.67867346e-02
9.43873383e-05 3.03390864e-02 8.42640293e-04 1.25434971e-03
9.44193080e-02 5.65310977e-02 6.39531091e-02 2.26973579e-03
8.35783035e-03 2.63884431e-03 1.12579623e-03 9.93307563e-04
2.08154244e-10 9.51728667e-04 1.52432276e-02 2.70403661e-02]
Category counts: [ 0 0 0 173 419 2 0 49 0 515 8 18 560 115 14 407 73 0
52 595 19 125 21 278 0 154 3 7 494 260 330 15 41 10 9 5
0 4 101 124]
Current means: [ nan nan nan 44.21294 68.98006 48.244038 nan
48.57217 nan 76.39027 49.95064 49.21184 84.984344 56.27346
49.05907 30.137125 46.812183 nan 50.955772 14.281114 50.466686
59.81618 47.545956 63.32053 nan 53.97194 46.39124 48.35056
22.977398 40.802254 36.16964 52.023277 47.162632 48.391277 47.299946
50.282494 nan 47.178276 51.764126 58.43635 ]
Posterior variance: [1.0000000e+01 1.0000000e+01 1.0000000e+01 5.7770074e-03 2.3860654e-03
4.7619051e-01 1.0000000e+01 2.0366600e-02 1.0000000e+01 1.9413707e-03
1.2345678e-01 5.5248618e-02 1.7853956e-03 8.6880978e-03 7.0921987e-02
2.4563989e-03 1.3679891e-02 1.0000000e+01 1.9193858e-02 1.6803899e-03
5.2356020e-02 7.9936050e-03 4.7393363e-02 3.5958288e-03 1.0000000e+01
6.4892923e-03 3.2258067e-01 1.4084508e-01 2.0238818e-03 3.8446751e-03
3.0293849e-03 6.6225164e-02 2.4330901e-02 9.9009894e-02 1.0989010e-01
1.9607843e-01 1.0000000e+01 2.4390244e-01 9.8911971e-03 8.0580181e-03]
Posterior means: [ nan nan nan 44.216286 68.975525 48.327652 nan
48.575077 nan 76.38514 49.951252 49.216198 84.978096 56.268013
49.065746 30.142004 46.816547 nan 50.953938 14.287116 50.464245
59.80833 47.557583 63.31574 nan 53.969364 46.507652 48.373787
22.982866 40.80579 36.17383 52.00988 47.169537 48.4072 47.32962
50.276955 nan 47.2471 51.762383 58.429554]
Resampled means: [ nan nan nan 44.21199 68.977234 47.88106 nan
48.5615 nan 76.38218 50.009598 49.313316 84.9772 56.26208
49.079582 30.141233 46.814335 nan 50.998215 14.288943 50.601223
59.812035 47.580566 63.31232 nan 53.96958 46.331688 48.26493
22.9811 40.804367 36.18016 52.03869 47.169132 48.48611 47.36634
50.199207 nan 47.075287 51.74989 58.438393]
Chosen means: [54.785324 48.72506 50.256966 44.21199 68.977234 47.88106 51.656677
48.5615 47.863136 76.38218 50.009598 49.313316 84.9772 56.26208
49.079582 30.141233 46.814335 50.389515 50.998215 14.288943 50.601223
59.812035 47.580566 63.31232 55.074005 53.96958 46.331688 48.26493
22.9811 40.804367 36.18016 52.03869 47.169132 48.48611 47.36634
50.199207 51.840244 47.075287 51.74989 58.438393]
Category counts: [ 0 1 0 186 445 0 0 54 0 507 5 22 532 122 12 426 79 0
64 580 12 128 22 273 1 144 3 10 487 254 321 22 43 12 4 7
0 3 83 136]
New alpha: [3.125000e-01 1.312500e+00 3.125000e-01 1.863125e+02 4.453125e+02
3.125000e-01 3.125000e-01 5.431250e+01 3.125000e-01 5.073125e+02
5.312500e+00 2.231250e+01 5.323125e+02 1.223125e+02 1.231250e+01
4.263125e+02 7.931250e+01 3.125000e-01 6.431250e+01 5.803125e+02
1.231250e+01 1.283125e+02 2.231250e+01 2.733125e+02 1.312500e+00
1.443125e+02 3.312500e+00 1.031250e+01 4.873125e+02 2.543125e+02
3.213125e+02 2.231250e+01 4.331250e+01 1.231250e+01 4.312500e+00
7.312500e+00 3.125000e-01 3.312500e+00 8.331250e+01 1.363125e+02]
New probs: [2.80429322e-06 1.81943644e-04 9.73285933e-06 3.57217528e-02
8.98862183e-02 1.03743474e-07 7.27040315e-05 1.15019511e-02
3.98485921e-04 1.03315681e-01 7.49598083e-04 2.75973906e-03
1.06978238e-01 2.51188092e-02 1.92470034e-03 8.18696171e-02
1.42404065e-02 6.34031494e-06 1.47428839e-02 1.16038099e-01
2.59371079e-03 2.93435864e-02 3.47060990e-03 5.51026911e-02
3.09708557e-04 2.99854819e-02 4.79521579e-04 2.44962028e-03
9.40896422e-02 5.23621067e-02 6.60327673e-02 3.30943568e-03
9.25188418e-03 3.47817922e-03 6.33124553e-04 1.21755700e-03
9.23704574e-05 3.65160522e-04 1.41465487e-02 2.57662330e-02]
Category counts: [ 0 1 0 186 445 0 0 54 0 507 5 22 532 122 12 426 79 0
64 580 12 128 22 273 1 144 3 10 487 254 321 22 43 12 4 7
0 3 83 136]
Current means: [ nan 48.224773 nan 44.049427 69.33452 nan nan
48.586113 nan 76.88929 50.48852 49.562607 85.22019 56.170803
48.88553 29.94153 46.61958 nan 50.90063 14.166531 51.151443
60.042255 47.853283 63.549507 54.080788 53.989964 46.77264 48.215797
22.67945 40.57391 36.037235 52.287537 46.94919 48.87729 47.423836
50.096897 nan 47.220306 51.81302 58.468895]
Posterior variance: [1.00000000e+01 9.09090877e-01 1.00000000e+01 5.37345512e-03
2.24668602e-03 1.00000000e+01 1.00000000e+01 1.84842888e-02
1.00000000e+01 1.97199755e-03 1.96078435e-01 4.52488698e-02
1.87934609e-03 8.19000788e-03 8.26446265e-02 2.34686700e-03
1.26422253e-02 1.00000000e+01 1.56006245e-02 1.72384083e-03
8.26446265e-02 7.80640077e-03 4.52488698e-02 3.66166234e-03
9.09090877e-01 6.93962490e-03 3.22580665e-01 9.90098938e-02
2.05296651e-03 3.93545860e-03 3.11429449e-03 4.52488698e-02
2.32018568e-02 8.26446265e-02 2.43902445e-01 1.40845075e-01
1.00000000e+01 3.22580665e-01 1.20336944e-02 7.34753814e-03]
Posterior means: [ nan 48.38616 nan 44.052624 69.33017 nan nan
48.58873 nan 76.88399 50.478943 49.564583 85.21358 56.16575
48.89474 29.946238 46.62385 nan 50.899223 14.172708 51.14193
60.034416 47.862995 63.544544 53.70981 53.987194 46.87675 48.23346
22.685059 40.57762 36.04158 52.277184 46.95627 48.88657 47.48667
50.09553 nan 47.309975 51.81084 58.462673]
Resampled means: [ nan 48.517757 nan 44.050083 69.32965 nan nan
48.58809 nan 76.8844 50.43929 49.565693 85.21376 56.158504
48.87032 29.946695 46.607834 nan 50.90474 14.172882 51.01044
60.040585 47.870296 63.54323 52.854137 53.988655 46.597153 48.191963
22.686249 40.57778 36.043633 52.25087 46.963257 48.82206 47.190994
50.283943 nan 47.756027 51.807682 58.463276]
Chosen means: [54.785324 48.517757 50.256966 44.050083 69.32965 47.88106 51.656677
48.58809 47.863136 76.8844 50.43929 49.565693 85.21376 56.158504
48.87032 29.946695 46.607834 50.389515 50.90474 14.172882 51.01044
60.040585 47.870296 63.54323 52.854137 53.988655 46.597153 48.191963
22.686249 40.57778 36.043633 52.25087 46.963257 48.82206 47.190994
50.283943 51.840244 47.756027 51.807682 58.463276]
Category counts: [ 0 0 0 187 441 0 0 63 4 504 7 6 502 122 8 453 75 0
77 562 19 154 20 293 1 171 1 10 475 267 305 14 49 15 2 7
0 2 67 117]
New alpha: [3.125000e-01 3.125000e-01 3.125000e-01 1.873125e+02 4.413125e+02
3.125000e-01 3.125000e-01 6.331250e+01 4.312500e+00 5.043125e+02
7.312500e+00 6.312500e+00 5.023125e+02 1.223125e+02 8.312500e+00
4.533125e+02 7.531250e+01 3.125000e-01 7.731250e+01 5.623125e+02
1.931250e+01 1.543125e+02 2.031250e+01 2.933125e+02 1.312500e+00
1.713125e+02 1.312500e+00 1.031250e+01 4.753125e+02 2.673125e+02
3.053125e+02 1.431250e+01 4.931250e+01 1.531250e+01 2.312500e+00
7.312500e+00 3.125000e-01 2.312500e+00 6.731250e+01 1.173125e+02]
New probs: [7.09650340e-06 1.83198736e-05 1.14397597e-04 3.48398425e-02
8.63475204e-02 1.40617340e-05 7.28057756e-08 1.07375896e-02
1.21475896e-03 1.04338519e-01 1.15982268e-03 8.00847600e-04
1.06258914e-01 2.34691277e-02 2.06163968e-03 9.13328081e-02
1.59997996e-02 1.57997158e-04 1.35226781e-02 1.05034478e-01
3.22956499e-03 2.88475659e-02 4.06102045e-03 6.06048405e-02
4.57637972e-04 3.27504836e-02 2.46213749e-05 1.79761450e-03
9.41155851e-02 5.56505769e-02 6.28353655e-02 1.77561468e-03
8.04054551e-03 2.61475542e-03 1.49044383e-03 1.19895034e-03
1.66104379e-04 9.06577450e-04 1.62409842e-02 2.57608406e-02]
Category counts: [ 0 0 0 187 441 0 0 63 4 504 7 6 502 122 8 453 75 0
77 562 19 154 20 293 1 171 1 10 475 267 305 14 49 15 2 7
0 2 67 117]
Current means: [ nan nan nan 43.923756 69.83612 nan nan
48.509785 47.89221 77.3927 49.40949 48.98676 85.48665 56.266586
48.468006 29.686811 46.63543 nan 50.841892 14.025418 50.821384
60.23792 47.7693 64.026245 52.705917 53.939037 46.64162 48.106464
22.28939 40.37375 35.860214 52.296173 46.936718 48.68881 47.494473
50.590763 nan 48.45289 51.737633 58.453934]
Posterior variance: [1.0000000e+01 1.0000000e+01 1.0000000e+01 5.3447355e-03 2.2670596e-03
1.0000000e+01 1.0000000e+01 1.5847862e-02 2.4390244e-01 1.9837334e-03
1.4084508e-01 1.6393442e-01 1.9916352e-03 8.1900079e-03 1.2345678e-01
2.2070182e-03 1.3315580e-02 1.0000000e+01 1.2970169e-02 1.7790430e-03
5.2356020e-02 6.4892923e-03 4.9751244e-02 3.4118048e-03 9.0909088e-01
5.8445353e-03 9.0909088e-01 9.9009894e-02 2.1048200e-03 3.7439161e-03
3.2776138e-03 7.0921987e-02 2.0366600e-02 6.6225164e-02 4.7619051e-01
1.4084508e-01 1.0000000e+01 4.7619051e-01 1.4903130e-02 8.5397102e-03]
Posterior means: [ nan nan nan 43.927006 69.83163 nan nan
48.512146 47.94362 77.38726 49.417805 49.00337 85.479576 56.26145
48.48692 29.691296 46.639908 nan 50.840797 14.031817 50.817085
60.231277 47.780396 64.02146 52.459927 53.936737 46.94693 48.12521
22.295223 40.377354 35.86485 52.27989 46.942955 48.697495 47.61378
50.58244 nan 48.526558 51.735046 58.446716]
Resampled means: [ nan nan nan 43.926414 69.83049 nan nan
48.483227 47.84457 77.38935 49.23747 49.086277 85.48034 56.252865
48.56688 29.686947 46.626675 nan 50.84313 14.033538 50.946392
60.22171 47.765667 64.02113 52.524822 53.93591 46.681522 48.02826
22.294777 40.375267 35.86552 52.230457 46.93764 48.691357 48.026367
50.673035 nan 49.046455 51.741783 58.453773]
Chosen means: [54.785324 48.517757 50.256966 43.926414 69.83049 47.88106 51.656677
48.483227 47.84457 77.38935 49.23747 49.086277 85.48034 56.252865
48.56688 29.686947 46.626675 50.389515 50.84313 14.033538 50.946392
60.22171 47.765667 64.02113 52.524822 53.93591 46.681522 48.02826
22.294777 40.375267 35.86552 52.230457 46.93764 48.691357 48.026367
50.673035 51.840244 49.046455 51.741783 58.453773]
Category counts: [ 0 0 0 203 400 0 0 44 13 500 12 8 500 130 12 455 81 2
75 527 15 146 19 333 3 155 0 5 485 263 313 10 47 12 4 4
0 8 87 129]
New alpha: [3.125000e-01 3.125000e-01 3.125000e-01 2.033125e+02 4.003125e+02
3.125000e-01 3.125000e-01 4.431250e+01 1.331250e+01 5.003125e+02
1.231250e+01 8.312500e+00 5.003125e+02 1.303125e+02 1.231250e+01
4.553125e+02 8.131250e+01 2.312500e+00 7.531250e+01 5.273125e+02
1.531250e+01 1.463125e+02 1.931250e+01 3.333125e+02 3.312500e+00
1.553125e+02 3.125000e-01 5.312500e+00 4.853125e+02 2.633125e+02
3.133125e+02 1.031250e+01 4.731250e+01 1.231250e+01 4.312500e+00
4.312500e+00 3.125000e-01 8.312500e+00 8.731250e+01 1.293125e+02]
New probs: [4.9949818e-13 1.4943035e-07 6.6575382e-08 4.1952372e-02 8.5524663e-02
5.1285926e-04 6.4538490e-09 7.7770674e-03 3.7312214e-03 1.0055984e-01
2.7382523e-03 2.0116167e-03 1.0221276e-01 2.4806760e-02 1.7664705e-03
9.6070290e-02 1.6504044e-02 4.3739934e-04 1.8986214e-02 9.9815868e-02
2.7849060e-03 2.8823545e-02 3.4886771e-03 6.0641844e-02 8.4518711e-04
3.2398392e-02 3.7109229e-04 9.1513933e-04 1.0045167e-01 5.1131386e-02
5.7238638e-02 1.6108911e-03 9.6134301e-03 1.2128513e-03 2.8153794e-04
6.0813525e-04 6.9596456e-07 1.8938705e-03 1.6075103e-02 2.4205066e-02]
Category counts: [ 0 0 0 203 400 0 0 44 13 500 12 8 500 130 12 455 81 2
75 527 15 146 19 333 3 155 0 5 485 263 313 10 47 12 4 4
0 8 87 129]
Current means: [ nan nan nan 43.779316 70.27104 nan nan
48.43407 47.824604 77.4616 49.09482 49.27849 85.505264 56.342457
48.087402 29.345306 46.639038 50.407413 50.927307 13.741527 50.573505
60.41481 47.504684 64.4172 51.702713 53.98133 nan 47.838524
21.79433 40.10696 35.507427 52.560955 47.005585 48.646725 48.266712
50.46382 nan 49.48398 51.709164 58.534435]
Posterior variance: [1.0000000e+01 1.0000000e+01 1.0000000e+01 4.9236827e-03 2.4993750e-03
1.0000000e+01 1.0000000e+01 2.2675738e-02 7.6335877e-02 1.9996001e-03
8.2644626e-02 1.2345678e-01 1.9996001e-03 7.6863947e-03 8.2644626e-02
2.1973192e-03 1.2330457e-02 4.7619051e-01 1.3315580e-02 1.8971734e-03
6.6225164e-02 6.8446267e-03 5.2356020e-02 3.0021013e-03 3.2258067e-01
6.4474531e-03 1.0000000e+01 1.9607843e-01 2.0614306e-03 3.8008362e-03
3.1938676e-03 9.9009894e-02 2.1231424e-02 8.2644626e-02 2.4390244e-01
2.4390244e-01 1.0000000e+01 1.2345678e-01 1.1481057e-02 7.7459328e-03]
Posterior means: [ nan nan nan 43.782383 70.265976 nan nan
48.437626 47.84121 77.45611 49.102303 49.287395 85.49816 56.33758
48.10321 29.349844 46.643185 50.388012 50.92607 13.748405 50.56971
60.40768 47.51775 64.412865 51.64779 53.978764 nan 47.880905
21.800146 40.11072 35.512054 52.535595 47.01194 48.65791 48.30899
50.45251 nan 49.49035 51.707195 58.52783 ]
Resampled means: [ nan nan nan 43.783382 70.26693 nan nan
48.45525 47.874165 77.45567 49.03182 49.291134 85.49722 56.334892
48.101994 29.352394 46.643364 50.022465 50.92127 13.748005 50.57531
60.399902 47.525337 64.41111 51.772026 53.974533 nan 47.608585
21.801092 40.106842 35.51583 52.671757 46.97978 48.666348 48.26028
50.652542 nan 49.535423 51.70479 58.538273]
Chosen means: [54.785324 48.517757 50.256966 43.783382 70.26693 47.88106 51.656677
48.45525 47.874165 77.45567 49.03182 49.291134 85.49722 56.334892
48.101994 29.352394 46.643364 50.022465 50.92127 13.748005 50.57531
60.399902 47.525337 64.41111 51.772026 53.974533 46.681522 47.608585
21.801092 40.106842 35.51583 52.671757 46.97978 48.666348 48.26028
50.652542 51.840244 49.535423 51.70479 58.538273]
Category counts: [ 0 0 0 214 381 3 0 44 22 496 13 9 499 141 8 442 86 2
84 501 16 167 16 338 6 153 1 3 499 272 314 8 52 3 1 4
0 7 70 125]
New alpha: [3.125000e-01 3.125000e-01 3.125000e-01 2.143125e+02 3.813125e+02
3.312500e+00 3.125000e-01 4.431250e+01 2.231250e+01 4.963125e+02
1.331250e+01 9.312500e+00 4.993125e+02 1.413125e+02 8.312500e+00
4.423125e+02 8.631250e+01 2.312500e+00 8.431250e+01 5.013125e+02
1.631250e+01 1.673125e+02 1.631250e+01 3.383125e+02 6.312500e+00
1.533125e+02 1.312500e+00 3.312500e+00 4.993125e+02 2.723125e+02
3.143125e+02 8.312500e+00 5.231250e+01 3.312500e+00 1.312500e+00
4.312500e+00 3.125000e-01 7.312500e+00 7.031250e+01 1.253125e+02]
New probs: [5.5907103e-06 6.2187646e-05 6.8134519e-05 5.0386868e-02 7.2158583e-02
2.2513649e-04 2.0889744e-04 6.7379922e-03 4.0450436e-03 9.8244518e-02
3.5109459e-03 2.5528576e-03 9.8781064e-02 2.8623892e-02 1.2686024e-03
8.6860046e-02 1.4883361e-02 5.6168222e-04 1.4298958e-02 1.0015413e-01
3.1936751e-03 3.4063544e-02 3.6306994e-03 6.6614963e-02 9.0740394e-04
3.2160938e-02 1.2642148e-04 7.3079928e-04 9.9814013e-02 5.3025357e-02
6.3014135e-02 1.5182944e-03 1.1054564e-02 8.1649300e-04 8.0712831e-05
2.7004543e-03 9.1134338e-05 1.9154303e-03 1.4909351e-02 2.5993394e-02]
Category counts: [ 0 0 0 214 381 3 0 44 22 496 13 9 499 141 8 442 86 2
84 501 16 167 16 338 6 153 1 3 499 272 314 8 52 3 1 4
0 7 70 125]
Current means: [ nan nan nan 43.608616 70.54277 48.219234 nan
48.420963 48.1218 77.50534 49.01728 49.216335 85.51192 56.252937
48.62314 29.072632 46.672 50.668213 50.92214 13.509985 50.57722
60.67238 47.786343 64.70131 52.178806 53.86198 46.042553 48.840153
21.503424 39.84333 35.060436 52.554306 46.82122 48.492023 48.865356
51.927937 nan 49.3791 51.670097 58.465515]
Posterior variance: [1.0000000e+01 1.0000000e+01 1.0000000e+01 4.6707145e-03 2.6239832e-03
3.2258067e-01 1.0000000e+01 2.2675738e-02 4.5248870e-02 2.0157227e-03
7.6335877e-02 1.0989010e-01 2.0036064e-03 7.0871720e-03 1.2345678e-01
2.2619315e-03 1.1614402e-02 4.7619051e-01 1.1890607e-02 1.9956096e-03
6.2111799e-02 5.9844404e-03 6.2111799e-02 2.9577047e-03 1.6393442e-01
6.5316786e-03 9.0909088e-01 3.2258067e-01 2.0036064e-03 3.6751193e-03
3.1836994e-03 1.2345678e-01 1.9193858e-02 3.2258067e-01 9.0909088e-01
2.4390244e-01 1.0000000e+01 1.4084508e-01 1.4265335e-02 7.9936050e-03]
Posterior means: [ nan nan nan 43.611603 70.537384 48.27668 nan
48.42455 48.1303 77.499794 49.02478 49.22495 85.50481 56.2485
48.640137 29.077368 46.675865 50.63639 50.92104 13.517267 50.573635
60.665993 47.80009 64.69696 52.143085 53.85946 46.402325 48.877567
21.509132 39.84706 35.065193 52.52277 46.827324 48.54067 48.96851
51.880917 nan 49.387844 51.667713 58.458744]
Resampled means: [ nan nan nan 43.617302 70.53343 48.199238 nan
48.38795 48.209743 77.495544 48.96209 49.337048 85.50465 56.244225
48.624043 29.078491 46.678635 50.71531 50.905518 13.516542 50.530838
60.665016 47.87274 64.70049 52.353455 53.86357 46.48974 48.60417
21.50624 39.845234 35.066395 52.49889 46.828976 48.101315 50.099075
52.108906 nan 49.385258 51.663452 58.459953]
Chosen means: [54.785324 48.517757 50.256966 43.617302 70.53343 48.199238 51.656677
48.38795 48.209743 77.495544 48.96209 49.337048 85.50465 56.244225
48.624043 29.078491 46.678635 50.71531 50.905518 13.516542 50.530838
60.665016 47.87274 64.70049 52.353455 53.86357 46.48974 48.60417
21.50624 39.845234 35.066395 52.49889 46.828976 48.101315 50.099075
52.108906 51.840244 49.385258 51.663452 58.459953]
Category counts: [ 0 0 0 240 384 0 1 30 29 487 23 11 501 144 7 406 80 6
63 501 25 192 21 312 3 143 0 2 495 264 344 3 49 3 0 17
0 11 72 131]
New alpha: [3.125000e-01 3.125000e-01 3.125000e-01 2.403125e+02 3.843125e+02
3.125000e-01 1.312500e+00 3.031250e+01 2.931250e+01 4.873125e+02
2.331250e+01 1.131250e+01 5.013125e+02 1.443125e+02 7.312500e+00
4.063125e+02 8.031250e+01 6.312500e+00 6.331250e+01 5.013125e+02
2.531250e+01 1.923125e+02 2.131250e+01 3.123125e+02 3.312500e+00
1.433125e+02 3.125000e-01 2.312500e+00 4.953125e+02 2.643125e+02
3.443125e+02 3.312500e+00 4.931250e+01 3.312500e+00 3.125000e-01
1.731250e+01 3.125000e-01 1.131250e+01 7.231250e+01 1.313125e+02]
New probs: [3.25791291e-07 2.44681132e-06 3.21692496e-05 4.82545905e-02
8.52804482e-02 5.00553824e-06 6.75006362e-04 7.38058286e-03
5.84495906e-03 9.83613282e-02 6.14352431e-03 2.03070883e-03
1.01843886e-01 2.72703320e-02 1.30591972e-03 8.15449506e-02
1.78689770e-02 7.00326636e-04 1.33592850e-02 1.00683339e-01
4.42511821e-03 3.87859121e-02 4.76672594e-03 6.19496927e-02
6.13622251e-04 2.65059173e-02 4.54518304e-05 6.45950553e-04
9.37550068e-02 4.88904938e-02 6.89235106e-02 4.20088589e-04
8.45643878e-03 5.90828247e-04 1.62154901e-10 2.93524005e-03
2.89764248e-05 2.52371025e-03 1.14084650e-02 2.57404931e-02]
Category counts: [ 0 0 0 240 384 0 1 30 29 487 23 11 501 144 7 406 80 6
63 501 25 192 21 312 3 143 0 2 495 264 344 3 49 3 0 17
0 11 72 131]
Current means: [ nan nan nan 43.60038 70.63417 nan 50.72257
48.519573 48.361176 77.536606 49.13186 49.153957 85.49552 56.21049
48.2487 28.704706 46.71353 51.225296 50.92542 13.50986 50.689518
61.03122 47.996216 64.96309 52.849346 53.903244 nan 48.337147
21.47621 39.56952 34.681812 52.485977 46.724266 47.1351 nan
51.934 nan 49.618458 51.730812 58.41097 ]
Posterior variance: [1.00000000e+01 1.00000000e+01 1.00000000e+01 4.16493136e-03
2.60348874e-03 1.00000000e+01 9.09090877e-01 3.32225896e-02
3.43642607e-02 2.05296651e-03 4.32900414e-02 9.00900885e-02
1.99560961e-03 6.93962490e-03 1.40845075e-01 2.46244762e-03
1.24843949e-02 1.63934425e-01 1.58478618e-02 1.99560961e-03
3.98406386e-02 5.20562194e-03 4.73933630e-02 3.20410123e-03
3.22580665e-01 6.98811980e-03 1.00000000e+01 4.76190507e-01
2.01979396e-03 3.78644443e-03 2.90613179e-03 3.22580665e-01
2.03665998e-02 3.22580665e-01 1.00000000e+01 5.84795326e-02
1.00000000e+01 9.00900885e-02 1.38696255e-02 7.62776472e-03]
Posterior means: [ nan nan nan 43.603046 70.6288 nan 50.656883
48.52449 48.366806 77.53095 49.13562 49.16158 85.48844 56.20618
48.273365 28.70995 46.717636 51.205208 50.923954 13.517142 50.68677
61.02548 48.00571 64.9583 52.757435 53.900513 nan 48.41633
21.481972 39.573467 34.686264 52.405785 46.730938 47.22752 nan
51.922688 nan 49.621895 51.72841 58.404556]
Resampled means: [ nan nan nan 43.605865 70.62407 nan 48.865368
48.527267 48.334667 77.531235 49.14076 49.17936 85.48706 56.189693
48.351044 28.709566 46.733047 51.08135 50.917896 13.517806 50.63491
61.030266 48.03768 64.95566 52.91168 53.90314 nan 48.68394
21.480673 39.572712 34.68572 52.473522 46.724045 47.082855 nan
51.952805 nan 49.749027 51.72708 58.402912]
Chosen means: [54.785324 48.517757 50.256966 43.605865 70.62407 48.199238 48.865368
48.527267 48.334667 77.531235 49.14076 49.17936 85.48706 56.189693
48.351044 28.709566 46.733047 51.08135 50.917896 13.517806 50.63491
61.030266 48.03768 64.95566 52.91168 53.90314 46.48974 48.68394
21.480673 39.572712 34.68572 52.473522 46.724045 47.082855 50.099075
51.952805 51.840244 49.749027 51.72708 58.402912]
Category counts: [ 0 0 0 238 388 0 3 40 34 483 23 13 499 143 6 383 91 4
65 501 26 218 18 279 6 145 1 4 495 260 363 4 35 7 0 11
0 10 63 141]
New alpha: [3.125000e-01 3.125000e-01 3.125000e-01 2.383125e+02 3.883125e+02
3.125000e-01 3.312500e+00 4.031250e+01 3.431250e+01 4.833125e+02
2.331250e+01 1.331250e+01 4.993125e+02 1.433125e+02 6.312500e+00
3.833125e+02 9.131250e+01 4.312500e+00 6.531250e+01 5.013125e+02
2.631250e+01 2.183125e+02 1.831250e+01 2.793125e+02 6.312500e+00
1.453125e+02 1.312500e+00 4.312500e+00 4.953125e+02 2.603125e+02
3.633125e+02 4.312500e+00 3.531250e+01 7.312500e+00 3.125000e-01
1.131250e+01 3.125000e-01 1.031250e+01 6.331250e+01 1.413125e+02]
New probs: [3.4635155e-05 2.8219031e-06 9.7406832e-05 4.7968537e-02 8.0768242e-02
3.0099840e-05 5.5328326e-04 7.6829586e-03 6.3383202e-03 9.1058545e-02
4.5402637e-03 2.1751572e-03 1.0221734e-01 2.4443300e-02 1.2629405e-03
7.9693839e-02 1.9294446e-02 1.3136385e-03 1.1645088e-02 9.8354578e-02
3.7392078e-03 4.5674659e-02 4.1715070e-03 5.4907314e-02 1.2310073e-03
2.8054828e-02 5.8268208e-05 7.8204012e-04 9.6263528e-02 5.3029835e-02
7.9439983e-02 2.0693346e-04 6.9700917e-03 1.0893432e-03 9.5995591e-04
1.6939358e-03 1.3488378e-05 2.0413008e-03 1.1068141e-02 2.9129606e-02]
Category counts: [ 0 0 0 238 388 0 3 40 34 483 23 13 499 143 6 383 91 4
65 501 26 218 18 279 6 145 1 4 495 260 363 4 35 7 0 11
0 10 63 141]
Current means: [ nan nan nan 43.47454 70.70658 nan 49.58031
48.555317 48.17765 77.59376 49.263752 49.05468 85.51204 56.19612
47.46115 28.490477 46.593445 51.40793 51.00254 13.509324 50.930656
61.38901 47.733627 65.25762 52.106464 53.848 48.241695 49.778008
21.477484 39.46377 34.501064 52.70704 46.481365 47.326714 nan
51.93296 nan 49.75189 51.796474 58.47957 ]
Posterior variance: [1.0000000e+01 1.0000000e+01 1.0000000e+01 4.1999160e-03 2.5766555e-03
1.0000000e+01 3.2258067e-01 2.4937658e-02 2.9325515e-02 2.0699648e-03
4.3290041e-02 7.6335877e-02 2.0036064e-03 6.9881198e-03 1.6393442e-01
2.6102844e-03 1.0976949e-02 2.4390244e-01 1.5360983e-02 1.9956096e-03
3.8314175e-02 4.5850524e-03 5.5248618e-02 3.5829451e-03 1.6393442e-01
6.8917987e-03 9.0909088e-01 2.4390244e-01 2.0197940e-03 3.8446751e-03
2.7540622e-03 2.4390244e-01 2.8490029e-02 1.4084508e-01 1.0000000e+01
9.0090089e-02 1.0000000e+01 9.9009894e-02 1.5847862e-02 7.0871720e-03]
Posterior means: [ nan nan nan 43.47728 70.70124 nan 49.59385
48.558918 48.183 77.58805 49.26694 49.061893 85.50493 56.191788
47.50277 28.49609 46.597183 51.373592 51.001 13.516606 50.92709
61.383793 47.74615 65.25215 52.07193 53.845345 48.401543 49.783424
21.483246 39.467823 34.505333 52.641014 46.49139 47.364365 nan
51.915546 nan 49.75434 51.79363 58.473557]
Resampled means: [ nan nan nan 43.47686 70.7027 nan
49.827885 48.59339 48.160206 77.58965 49.281433 49.194508
85.50393 56.186115 47.539604 28.497272 46.58078 51.442986
50.97834 13.5168495 50.928905 61.39363 47.71127 65.249214
52.279945 53.8373 49.04661 49.99173 21.482107 39.46611
34.50521 52.413742 46.48437 47.582157 nan 51.891113
nan 49.85986 51.811466 58.478897 ]
Chosen means: [54.785324 48.517757 50.256966 43.47686 70.7027 48.199238
49.827885 48.59339 48.160206 77.58965 49.281433 49.194508
85.50393 56.186115 47.539604 28.497272 46.58078 51.442986
50.97834 13.5168495 50.928905 61.39363 47.71127 65.249214
52.279945 53.8373 49.04661 49.99173 21.482107 39.46611
34.50521 52.413742 46.48437 47.582157 50.099075 51.891113
51.840244 49.85986 51.811466 58.478897 ]
Category counts: [ 0 0 0 239 391 1 4 42 35 477 30 10 500 146 12 387 95 8
56 500 17 242 17 264 7 143 0 7 486 261 362 1 31 4 4 9
0 10 73 129]
New alpha: [3.125000e-01 3.125000e-01 3.125000e-01 2.393125e+02 3.913125e+02
1.312500e+00 4.312500e+00 4.231250e+01 3.531250e+01 4.773125e+02
3.031250e+01 1.031250e+01 5.003125e+02 1.463125e+02 1.231250e+01
3.873125e+02 9.531250e+01 8.312500e+00 5.631250e+01 5.003125e+02
1.731250e+01 2.423125e+02 1.731250e+01 2.643125e+02 7.312500e+00
1.433125e+02 3.125000e-01 7.312500e+00 4.863125e+02 2.613125e+02
3.623125e+02 1.312500e+00 3.131250e+01 4.312500e+00 4.312500e+00
9.312500e+00 3.125000e-01 1.031250e+01 7.331250e+01 1.293125e+02]
New probs: [3.0511830e-05 3.0462563e-04 1.4171136e-06 4.5171276e-02 7.0965245e-02
8.0417485e-05 8.3926145e-04 8.4822774e-03 6.2672999e-03 8.8894308e-02
5.7717143e-03 2.9643842e-03 1.0190670e-01 2.5496799e-02 1.8141996e-03
8.0302268e-02 1.6538337e-02 1.2334916e-03 1.1002403e-02 1.0415430e-01
2.3678588e-03 4.6080895e-02 3.9656893e-03 5.5080600e-02 5.3894758e-04
3.1780507e-02 6.1700397e-05 2.6583569e-03 9.8780356e-02 5.4839574e-02
7.3202141e-02 2.1951186e-05 6.7179217e-03 1.3174739e-03 5.2931532e-04
1.8078473e-03 4.1486866e-05 2.6748958e-03 1.9057769e-02 2.6253749e-02]
Category counts: [ 0 0 0 239 391 1 4 42 35 477 30 10 500 146 12 387 95 8
56 500 17 242 17 264 7 143 0 7 486 261 362 1 31 4 4 9
0 10 73 129]
Current means: [ nan nan nan 43.374447 70.75929 50.715942 48.891457
48.759766 47.982426 77.62051 49.14893 49.33862 85.505264 56.312946
47.39384 28.344835 46.528946 51.624245 51.096897 13.500216 50.6362
61.483482 47.80845 65.45166 51.58659 53.951187 nan 50.28064
21.402939 39.35617 34.404587 50.810223 46.44624 47.691452 50.625137
51.745853 nan 49.89942 51.78929 58.292572]
Posterior variance: [1.00000000e+01 1.00000000e+01 1.00000000e+01 4.18235036e-03
2.55689071e-03 9.09090877e-01 2.43902445e-01 2.37529706e-02
2.84900293e-02 2.09599664e-03 3.32225896e-02 9.90098938e-02
1.99960009e-03 6.84462674e-03 8.26446265e-02 2.58331187e-03
1.05152475e-02 1.23456784e-01 1.78253129e-02 1.99960009e-03
5.84795326e-02 4.13052458e-03 5.84795326e-02 3.78644443e-03
1.40845075e-01 6.98811980e-03 1.00000000e+01 1.40845075e-01
2.05718982e-03 3.82995023e-03 2.76166806e-03 9.09090877e-01
3.21543403e-02 2.43902445e-01 2.43902445e-01 1.09890103e-01
1.00000000e+01 9.90098938e-02 1.36798909e-02 7.74593279e-03]
Posterior means: [ nan nan nan 43.37722 70.75398 50.65086 48.918495
48.76271 47.988174 77.61472 49.151756 49.345165 85.49816 56.308624
47.415382 28.350431 46.532597 51.60419 51.09494 13.507514 50.63248
61.478737 47.821266 65.4458 51.56424 53.94842 nan 50.276684
21.408821 39.360245 34.408894 50.73657 46.457664 47.74776 50.609894
51.72667 nan 49.900414 51.786846 58.28615 ]
Resampled means: [ nan nan nan 43.38024 70.75457 50.042118 48.812843
48.755695 47.99515 77.613914 49.18082 49.475952 85.49539 56.309254
47.584938 28.352097 46.52681 51.462982 51.084747 13.509335 50.724102
61.485626 47.914547 65.44864 51.461273 53.940235 nan 50.01835
21.40581 39.35029 34.405388 50.99337 46.43412 47.726963 50.618633
51.7432 nan 50.12273 51.793316 58.286106]
Chosen means: [54.785324 48.517757 50.256966 43.38024 70.75457 50.042118 48.812843
48.755695 47.99515 77.613914 49.18082 49.475952 85.49539 56.309254
47.584938 28.352097 46.52681 51.462982 51.084747 13.509335 50.724102
61.485626 47.914547 65.44864 51.461273 53.940235 49.04661 50.01835
21.40581 39.35029 34.405388 50.99337 46.43412 47.726963 50.618633
51.7432 51.840244 50.12273 51.793316 58.286106]
Category counts: [ 0 1 0 233 378 0 5 34 43 476 35 10 499 124 10 394 88 4
53 499 13 245 13 270 3 158 0 10 482 279 346 0 35 6 2 8
0 16 81 147]
New alpha: [3.125000e-01 1.312500e+00 3.125000e-01 2.333125e+02 3.783125e+02
3.125000e-01 5.312500e+00 3.431250e+01 4.331250e+01 4.763125e+02
3.531250e+01 1.031250e+01 4.993125e+02 1.243125e+02 1.031250e+01
3.943125e+02 8.831250e+01 4.312500e+00 5.331250e+01 4.993125e+02
1.331250e+01 2.453125e+02 1.331250e+01 2.703125e+02 3.312500e+00
1.583125e+02 3.125000e-01 1.031250e+01 4.823125e+02 2.793125e+02
3.463125e+02 3.125000e-01 3.531250e+01 6.312500e+00 2.312500e+00
8.312500e+00 3.125000e-01 1.631250e+01 8.131250e+01 1.473125e+02]
New probs: [1.40619022e-06 4.74512752e-04 1.59367308e-04 4.18216661e-02
7.45415688e-02 8.56891447e-09 1.27747236e-03 8.71269405e-03
6.82738516e-03 9.82691646e-02 9.60949808e-03 2.10076664e-03
1.02366745e-01 2.44948957e-02 2.10511149e-03 7.59700239e-02
1.73265487e-02 2.41883346e-04 1.25310859e-02 1.01850301e-01
2.18132068e-03 4.61555123e-02 1.54434261e-03 5.70653155e-02
7.88126781e-04 2.88868379e-02 4.25003564e-05 3.09779611e-03
9.96703357e-02 5.31056225e-02 6.63886443e-02 6.45245236e-05
7.87119474e-03 8.89332674e-04 7.88078294e-04 2.39211298e-03
8.10353754e-07 3.60024697e-03 1.64868105e-02 2.82984544e-02]
Category counts: [ 0 1 0 233 378 0 5 34 43 476 35 10 499 124 10 394 88 4
53 499 13 245 13 270 3 158 0 10 482 279 346 0 35 6 2 8
0 16 81 147]
Current means: [ nan 50.236065 nan 43.39756 70.87479 nan 48.07522
48.63822 47.90253 77.64544 49.04522 49.846497 85.512146 56.357567
47.0487 28.314146 46.558346 51.39997 51.20949 13.493739 50.945477
61.570293 48.086613 65.683044 50.867588 53.947598 nan 49.707726
21.356606 39.257587 34.328957 nan 46.18873 47.523926 50.42093
51.829453 nan 49.840355 51.712902 58.300385]
Posterior variance: [1.0000000e+01 9.0909088e-01 1.0000000e+01 4.2900043e-03 2.6448029e-03
1.0000000e+01 1.9607843e-01 2.9325515e-02 2.3201857e-02 2.1003990e-03
2.8490029e-02 9.9009894e-02 2.0036064e-03 8.0580181e-03 9.9009894e-02
2.5374270e-03 1.1350738e-02 2.4390244e-01 1.8832393e-02 2.0036064e-03
7.6335877e-02 4.0799673e-03 7.6335877e-02 3.7023325e-03 3.2258067e-01
6.3251103e-03 1.0000000e+01 9.9009894e-02 2.0742584e-03 3.5829451e-03
2.8893382e-03 1.0000000e+01 2.8490029e-02 1.6393442e-01 4.7619051e-01
1.2345678e-01 1.0000000e+01 6.2111799e-02 1.2330457e-02 6.7980965e-03]
Posterior means: [ nan 50.214607 nan 43.400394 70.86926 nan 48.11296
48.64222 47.907402 77.63963 49.04794 49.84801 85.505035 56.352444
47.077915 28.319649 46.562252 51.365826 51.207214 13.501054 50.93826
61.565575 48.101215 65.67723 50.839603 53.9451 nan 49.710617
21.362549 39.261436 34.333485 nan 46.199585 47.564514 50.400883
51.806866 nan 49.841347 51.710793 58.294743]
Resampled means: [ nan 50.79293 nan 43.41106 70.86659 nan 48.36641
48.603058 47.89624 77.6405 49.02691 49.77925 85.50556 56.33808
47.03825 28.322071 46.58067 51.29116 51.208427 13.500567 50.91504
61.563812 48.169834 65.67794 50.623196 53.947845 nan 49.530136
21.361748 39.25987 34.331394 nan 46.282684 47.61319 50.91863
51.804142 nan 49.788506 51.721226 58.29742 ]
Chosen means: [54.785324 50.79293 50.256966 43.41106 70.86659 50.042118 48.36641
48.603058 47.89624 77.6405 49.02691 49.77925 85.50556 56.33808
47.03825 28.322071 46.58067 51.29116 51.208427 13.500567 50.91504
61.563812 48.169834 65.67794 50.623196 53.947845 49.04661 49.530136
21.361748 39.25987 34.331394 50.99337 46.282684 47.61319 50.91863
51.804142 51.840244 49.788506 51.721226 58.29742 ]
Category counts: [ 1 3 0 228 370 0 6 35 22 471 42 11 500 133 5 392 99 1
67 500 6 244 5 282 2 142 1 10 481 280 346 1 47 7 2 11
0 21 81 145]
New alpha: [1.312500e+00 3.312500e+00 3.125000e-01 2.283125e+02 3.703125e+02
3.125000e-01 6.312500e+00 3.531250e+01 2.231250e+01 4.713125e+02
4.231250e+01 1.131250e+01 5.003125e+02 1.333125e+02 5.312500e+00
3.923125e+02 9.931250e+01 1.312500e+00 6.731250e+01 5.003125e+02
6.312500e+00 2.443125e+02 5.312500e+00 2.823125e+02 2.312500e+00
1.423125e+02 1.312500e+00 1.031250e+01 4.813125e+02 2.803125e+02
3.463125e+02 1.312500e+00 4.731250e+01 7.312500e+00 2.312500e+00
1.131250e+01 3.125000e-01 2.131250e+01 8.131250e+01 1.453125e+02]
New probs: [2.43410395e-04 8.36246647e-04 2.96382859e-06 3.98644246e-02
7.71295577e-02 5.21818947e-05 1.31754111e-03 6.53072307e-03
4.90837870e-03 9.14968848e-02 8.20402708e-03 2.33462709e-03
9.96722803e-02 2.56021693e-02 9.38482583e-04 7.28163421e-02
2.28638500e-02 1.11488385e-04 1.31668076e-02 1.06436193e-01
8.38166045e-04 5.27748279e-02 9.27140820e-04 5.74082471e-02
3.52992822e-04 2.55355388e-02 1.35743539e-04 3.36077926e-03
9.80808809e-02 4.85726595e-02 6.85935766e-02 1.80162606e-04
9.87344421e-03 1.06530020e-03 1.56489783e-04 2.36505014e-03
3.21519019e-06 5.24714775e-03 1.60640068e-02 3.39357927e-02]
Category counts: [ 1 3 0 228 370 0 6 35 22 471 42 11 500 133 5 392 99 1
67 500 6 244 5 282 2 142 1 10 481 280 346 1 47 7 2 11
0 21 81 145]
Current means: [56.810204 50.38672 nan 43.361927 71.00005 nan 47.912918
48.528225 48.190784 77.665436 49.01738 50.496964 85.50273 56.169945
46.57189 28.296785 46.609184 52.158653 51.08198 13.500216 50.1309
61.572556 47.8621 65.794136 51.897995 54.058453 50.233772 49.086586
21.370134 39.22235 34.297623 50.55954 46.289062 47.340183 51.437
51.93249 nan 49.616375 51.789425 58.392456]
Posterior variance: [9.0909088e-01 3.2258067e-01 1.0000000e+01 4.3840418e-03 2.7019724e-03
1.0000000e+01 1.6393442e-01 2.8490029e-02 4.5248870e-02 2.1226916e-03
2.3752971e-02 9.0090089e-02 1.9996001e-03 7.5131478e-03 1.9607843e-01
2.5503698e-03 1.0090818e-02 9.0909088e-01 1.4903130e-02 1.9996001e-03
1.6393442e-01 4.0966817e-03 1.9607843e-01 3.5448421e-03 4.7619051e-01
7.0372974e-03 9.0909088e-01 9.9009894e-02 2.0785700e-03 3.5701534e-03
2.8893382e-03 9.0909088e-01 2.1231424e-02 1.4084508e-01 4.7619051e-01
9.0090089e-02 1.0000000e+01 4.7393363e-02 1.2330457e-02 6.8917987e-03]
Posterior means: [56.191097 50.374245 nan 43.364838 70.994385 nan 47.94713
48.532417 48.19897 77.65956 49.01971 50.49249 85.49563 56.16531
46.63911 28.30232 46.612602 51.962414 51.08037 13.507514 50.128754
61.56782 47.90402 65.788536 51.807613 54.0556 50.212524 49.095627
21.376085 39.226196 34.30216 50.508675 46.29694 47.377644 51.368572
51.91508 nan 49.61819 51.787224 58.38667 ]
Resampled means: [58.114204 50.720947 nan 43.3703 70.99138 nan 48.09725
48.512405 48.24592 77.66202 49.01609 50.587467 85.49391 56.167393
46.613876 28.300611 46.611897 52.33351 51.09628 13.509682 50.307884
61.5702 48.084915 65.78104 52.36102 54.05704 50.195206 49.23861
21.373919 39.228596 34.29975 50.328022 46.296467 47.315247 51.49997
51.87056 nan 49.583504 51.80251 58.386223]
Chosen means: [58.114204 50.720947 50.256966 43.3703 70.99138 50.042118 48.09725
48.512405 48.24592 77.66202 49.01609 50.587467 85.49391 56.167393
46.613876 28.300611 46.611897 52.33351 51.09628 13.509682 50.307884
61.5702 48.084915 65.78104 52.36102 54.05704 50.195206 49.23861
21.373919 39.228596 34.29975 50.328022 46.296467 47.315247 51.49997
51.87056 51.840244 49.583504 51.80251 58.386223]
Category counts: [ 3 4 0 225 376 0 8 37 25 454 43 15 500 130 6 385 107 0
61 500 8 248 3 288 1 142 0 14 483 279 348 0 45 6 1 11
0 26 72 146]
New alpha: [3.312500e+00 4.312500e+00 3.125000e-01 2.253125e+02 3.763125e+02
3.125000e-01 8.312500e+00 3.731250e+01 2.531250e+01 4.543125e+02
4.331250e+01 1.531250e+01 5.003125e+02 1.303125e+02 6.312500e+00
3.853125e+02 1.073125e+02 3.125000e-01 6.131250e+01 5.003125e+02
8.312500e+00 2.483125e+02 3.312500e+00 2.883125e+02 1.312500e+00
1.423125e+02 3.125000e-01 1.431250e+01 4.833125e+02 2.793125e+02
3.483125e+02 3.125000e-01 4.531250e+01 6.312500e+00 1.312500e+00
1.131250e+01 3.125000e-01 2.631250e+01 7.231250e+01 1.463125e+02]
New probs: [7.64535682e-04 6.84952596e-04 8.33735066e-06 4.51777205e-02
6.92145228e-02 1.20078628e-04 1.35255919e-03 6.58890605e-03
5.10420324e-03 9.34272185e-02 8.56507663e-03 4.58392967e-03
1.01584263e-01 2.62229331e-02 9.48858040e-04 8.16225633e-02
1.99896358e-02 4.38088546e-06 9.96625796e-03 9.54719558e-02
1.85955619e-03 5.08441553e-02 7.35376496e-04 6.06439821e-02
3.89003952e-04 2.65437216e-02 1.46178121e-04 2.13034637e-03
9.83637720e-02 5.36287911e-02 7.42327273e-02 2.08196838e-09
7.70636369e-03 1.95974112e-03 3.38508602e-04 2.39500613e-03
5.35312392e-05 5.77141158e-03 1.38524445e-02 2.70023812e-02]
Category counts: [ 3 4 0 225 376 0 8 37 25 454 43 15 500 130 6 385 107 0
61 500 8 248 3 288 1 142 0 14 483 279 348 0 45 6 1 11
0 26 72 146]
Current means: [57.7401 51.00595 nan 43.284855 71.22886 nan 47.59931
48.559303 48.208828 77.78626 49.237362 50.383904 85.505264 56.219234
45.831852 28.284973 46.411674 nan 51.488842 13.500216 50.520035
61.618484 48.65305 65.93817 54.100525 53.96384 nan 48.975773
21.382845 39.170288 34.22587 nan 46.43498 47.27304 52.271763
52.18589 nan 49.89228 51.831333 58.372726]
Posterior variance: [3.22580665e-01 2.43902445e-01 1.00000000e+01 4.44246968e-03
2.65886728e-03 1.00000000e+01 1.23456784e-01 2.69541796e-02
3.98406386e-02 2.20215810e-03 2.32018568e-02 6.62251636e-02
1.99960009e-03 7.68639473e-03 1.63934425e-01 2.59672804e-03
9.33706854e-03 1.00000000e+01 1.63666122e-02 1.99960009e-03
1.23456784e-01 4.03063279e-03 3.22580665e-01 3.47101688e-03
9.09090877e-01 7.03729736e-03 1.00000000e+01 7.09219873e-02
2.06996477e-03 3.58294509e-03 2.87273759e-03 1.00000000e+01
2.21729502e-02 1.63934425e-01 9.09090877e-01 9.00900885e-02
1.00000000e+01 3.83141749e-02 1.38696255e-02 6.84462674e-03]
Posterior means: [57.49042 50.98142 nan 43.287838 71.22322 nan 47.62895
48.563183 48.215965 77.78014 49.239136 50.381363 85.49816 56.214455
45.90018 28.290611 46.415024 nan 51.486404 13.507514 50.513615
61.613804 48.696503 65.93264 53.727753 53.961052 nan 48.98304
21.388767 39.174168 34.2304 nan 46.442883 47.31774 52.065243
52.1662 nan 49.892693 51.82879 58.366997]
Resampled means: [56.998425 50.860054 nan 43.280018 71.223854 nan 47.59126
48.565487 48.17486 77.77967 49.211807 50.30891 85.49646 56.19884
45.90986 28.292175 46.40763 nan 51.469707 13.507615 50.51911
61.618145 47.864704 65.92994 54.092648 53.954678 nan 49.02344
21.387598 39.172733 34.228004 nan 46.412285 47.197365 52.49266
52.111626 nan 49.916267 51.839035 58.371906]
Chosen means: [56.998425 50.860054 50.256966 43.280018 71.223854 50.042118 47.59126
48.565487 48.17486 77.77967 49.211807 50.30891 85.49646 56.19884
45.90986 28.292175 46.40763 52.33351 51.469707 13.507615 50.51911
61.618145 47.864704 65.92994 54.092648 53.954678 50.195206 49.02344
21.387598 39.172733 34.228004 50.328022 46.412285 47.197365 52.49266
52.111626 51.840244 49.916267 51.839035 58.371906]
Category counts: [ 6 3 0 225 377 0 4 33 31 439 40 20 499 137 5 389 108 0 49 499 16 248 5 301 5 138 1 8 482 292 336 0 41 12 1 11 1 30 67 141] New alpha: [6.312500e+00 3.312500e+00 3.125000e-01 2.253125e+02 3.773125e+02 3.125000e-01 4.312500e+00 3.331250e+01 3.131250e+01 4.393125e+02 4.031250e+01 2.031250e+01 4.993125e+02 1.373125e+02 5.312500e+00 3.893125e+02 1.083125e+02 3.125000e-01 4.931250e+01 4.993125e+02 1.631250e+01 2.483125e+02 5.312500e+00 3.013125e+02 5.312500e+00 1.383125e+02 1.312500e+00 8.312500e+00 4.823125e+02 2.923125e+02 3.363125e+02 3.125000e-01 4.131250e+01 1.231250e+01 1.312500e+00 1.131250e+01 1.312500e+00 3.031250e+01 6.731250e+01 1.413125e+02] New probs: [9.6678309e-04 3.9508296e-04 7.9507205e-05 4.7341980e-02 7.0675232e-02 1.2246127e-05 1.8909659e-03 5.2449168e-03 5.9702019e-03 8.6146668e-02 8.2107345e-03 3.0261944e-03 9.6346140e-02 2.8209724e-02 1.1206995e-03 7.7825800e-02 1.8845478e-02 1.4001704e-05 1.1767292e-02 1.0611850e-01 3.6945294e-03 5.2061781e-02 8.6982851e-04 6.5180548e-02 5.7372311e-04 2.8886149e-02 2.6476305e-04 1.2921882e-03 9.7034678e-02 5.0628766e-02 6.5972216e-02 3.1672324e-05 7.4010836e-03 2.7125934e-03 7.9258360e-05 1.6530894e-03 3.7436999e-04 5.7086516e-03 1.5905878e-02 2.9466076e-02] Category counts: [ 6 3 0 225 377 0 4 33 31 439 40 20 499 137 5 389 108 0 49 499 16 248 5 301 5 138 1 8 482 292 336 0 41 12 1 11 1 30 67 141] Current means: [56.644165 50.74463 nan 43.336224 71.47404 nan 47.2819 48.63369 48.129745 77.92116 49.187626 50.595375 85.512215 56.344917 45.034966 28.281582 46.500076 nan 51.368046 13.493734 50.587547 61.64951 48.19483 66.07235 54.69274 53.908894 48.90726 49.114185 21.35795 39.102074 34.173122 nan 46.354282 47.69212 54.222744 52.231934 50.904926 50.140934 51.950993 58.36867 ] Posterior variance: [1.6393442e-01 3.2258067e-01 1.0000000e+01 4.4424697e-03 2.6518165e-03 1.0000000e+01 2.4390244e-01 3.0211482e-02 3.2154340e-02 2.2773854e-03 2.4937658e-02 4.9751244e-02 2.0036064e-03 7.2939456e-03 1.9607843e-01 2.5700333e-03 9.2506940e-03 1.0000000e+01 2.0366600e-02 2.0036064e-03 6.2111799e-02 4.0306328e-03 1.9607843e-01 3.3211557e-03 1.9607843e-01 7.2411294e-03 9.0909088e-01 1.2345678e-01 2.0742584e-03 3.4234850e-03 2.9753048e-03 1.0000000e+01 2.4330901e-02 8.2644626e-02 9.0909088e-01 9.0090089e-02 9.0909088e-01 3.3222590e-02 1.4903130e-02 7.0871720e-03] Posterior means: [56.53524 50.72061 nan 43.339184 71.468346 nan 47.348194 48.63782 48.135757 77.914795 49.18965 50.59241 85.505104 56.34029 45.13232 28.287163 46.503315 nan 51.365257 13.501049 50.583897 61.644814 48.23023 66.06701 54.600727 53.906063 49.006603 49.125122 21.363894 39.105804 34.17783 nan 46.363155 47.711197 53.83886 52.211826 50.822662 50.140465 51.948086 58.362736] Resampled means: [56.413307 49.977238 nan 43.33512 71.472855 nan 47.28313 48.580143 48.18255 77.91594 49.162903 50.57166 85.5054 56.34658 45.05403 28.291475 46.491127 nan 51.389687 13.496277 50.63157 61.646587 48.401085 66.06896 54.81875 53.897095 49.89351 49.084488 21.367321 39.110237 34.179935 nan 46.38851 47.673836 53.050323 52.16431 50.84937 50.19548 51.93234 58.357082] Chosen means: [56.413307 49.977238 50.256966 43.33512 71.472855 50.042118 47.28313 48.580143 48.18255 77.91594 49.162903 50.57166 85.5054 56.34658 45.05403 28.291475 46.491127 52.33351 51.389687 13.496277 50.63157 61.646587 48.401085 66.06896 54.81875 53.897095 49.89351 49.084488 21.367321 39.110237 34.179935 50.328022 46.38851 47.673836 53.050323 52.16431 50.84937 50.19548 51.93234 58.357082] Category counts: [ 6 2 0 240 368 0 6 34 36 421 47 18 497 134 4 390 90 0 52 501 18 243 0 330 5 145 2 3 481 291 328 0 40 17 0 8 0 31 69 143] New alpha: [6.312500e+00 2.312500e+00 3.125000e-01 2.403125e+02 3.683125e+02 3.125000e-01 6.312500e+00 3.431250e+01 3.631250e+01 4.213125e+02 4.731250e+01 1.831250e+01 4.973125e+02 1.343125e+02 4.312500e+00 3.903125e+02 9.031250e+01 3.125000e-01 5.231250e+01 5.013125e+02 1.831250e+01 2.433125e+02 3.125000e-01 3.303125e+02 5.312500e+00 1.453125e+02 2.312500e+00 3.312500e+00 4.813125e+02 2.913125e+02 3.283125e+02 3.125000e-01 4.031250e+01 1.731250e+01 3.125000e-01 8.312500e+00 3.125000e-01 3.131250e+01 6.931250e+01 1.433125e+02] New probs: [6.9975248e-04 4.3831341e-04 3.5504536e-06 4.5476526e-02 7.3145725e-02 5.2649548e-06 1.8559048e-03 7.2659533e-03 7.9242876e-03 8.5444398e-02 8.3000055e-03 4.0000053e-03 1.0537560e-01 2.0916697e-02 1.1278843e-03 7.4036710e-02 1.8907314e-02 2.3818222e-07 9.2008635e-03 9.8215386e-02 2.9103516e-03 5.2052073e-02 5.0148341e-05 6.5897077e-02 7.7747792e-04 3.1257123e-02 5.5249786e-04 5.6265097e-04 9.6693248e-02 5.8242794e-02 6.5721817e-02 1.1008720e-04 9.1893012e-03 3.5699783e-03 2.0397223e-05 1.0119251e-03 1.2229506e-07 6.9588767e-03 1.2810428e-02 2.9271001e-02] Category counts: [ 6 2 0 240 368 0 6 34 36 421 47 18 497 134 4 390 90 0 52 501 18 243 0 330 5 145 2 3 481 291 328 0 40 17 0 8 0 31 69 143] Current means: [56.500458 50.115196 nan 43.33164 71.87142 nan 47.20531 48.50893 48.187103 78.10368 48.836655 50.78934 85.52416 56.34894 44.543365 28.30445 46.488434 nan 51.504326 13.510222 50.78502 61.683064 nan 66.292496 54.877052 53.935932 49.92327 49.860413 21.382296 38.99134 34.151592 nan 46.298588 47.83155 nan 52.664806 nan 50.380665 51.803783 58.491814] Posterior variance: [1.63934425e-01 4.76190507e-01 1.00000000e+01 4.16493136e-03 2.71665305e-03 1.00000000e+01 1.63934425e-01 2.93255150e-02 2.77008321e-02 2.37473287e-03 2.12314241e-02 5.52486181e-02 2.01166770e-03 7.45712128e-03 2.43902445e-01 2.56344513e-03 1.10987797e-02 1.00000000e+01 1.91938579e-02 1.99560961e-03 5.52486181e-02 4.11353353e-03 1.00000000e+01 3.02938488e-03 1.96078435e-01 6.89179869e-03 4.76190507e-01 3.22580665e-01 2.07856996e-03 3.43524548e-03 3.04785115e-03 1.00000000e+01 2.49376576e-02 5.84795326e-02 1.00000000e+01 1.23456784e-01 1.00000000e+01 3.21543403e-02 1.44717805e-02 6.98811980e-03] Posterior means: [56.39389 50.10971 nan 43.334415 71.86548 nan 47.25112 48.51331 48.192127 78.097 48.839123 50.78498 85.51701 56.344208 44.676456 28.31001 46.49233 nan 51.50144 13.517505 50.78068 61.67826 nan 66.28755 54.781425 53.933216 49.92692 49.864918 21.388245 38.995125 34.15642 nan 46.30782 47.84423 nan 52.63191 nan 50.37944 51.80117 58.485874] Resampled means: [56.538956 50.352234 nan 43.333897 71.86322 nan 47.411263 48.54042 48.14466 78.09408 48.858044 50.823296 85.51424 56.33515 44.689514 28.309277 46.484627 nan 51.482563 13.516183 50.785717 61.67584 nan 66.2892 55.02598 53.935493 50.394512 49.27578 21.385967 38.98882 34.155155 nan 46.301254 47.837715 nan 52.549976 nan 50.301365 51.801132 58.49296 ] Chosen means: [56.538956 50.352234 50.256966 43.333897 71.86322 50.042118 47.411263 48.54042 48.14466 78.09408 48.858044 50.823296 85.51424 56.33515 44.689514 28.309277 46.484627 52.33351 51.482563 13.516183 50.785717 61.67584 48.401085 66.2892 55.02598 53.935493 50.394512 49.27578 21.385967 38.98882 34.155155 50.328022 46.301254 47.837715 53.050323 52.549976 50.84937 50.301365 51.801132 58.49296 ] Category counts: [ 5 1 0 219 371 0 9 39 44 398 47 24 495 115 7 384 104 0 48 499 15 245 0 347 5 152 0 4 490 300 320 0 41 7 0 7 0 32 68 158] New alpha: [5.312500e+00 1.312500e+00 3.125000e-01 2.193125e+02 3.713125e+02 3.125000e-01 9.312500e+00 3.931250e+01 4.431250e+01 3.983125e+02 4.731250e+01 2.431250e+01 4.953125e+02 1.153125e+02 7.312500e+00 3.843125e+02 1.043125e+02 3.125000e-01 4.831250e+01 4.993125e+02 1.531250e+01 2.453125e+02 3.125000e-01 3.473125e+02 5.312500e+00 1.523125e+02 3.125000e-01 4.312500e+00 4.903125e+02 3.003125e+02 3.203125e+02 3.125000e-01 4.131250e+01 7.312500e+00 3.125000e-01 7.312500e+00 3.125000e-01 3.231250e+01 6.831250e+01 1.583125e+02] New probs: [6.14568940e-04 3.17575250e-05 1.77405236e-05 4.49006334e-02 7.04777539e-02 1.91179482e-04 1.27784652e-03 9.63288639e-03 9.42455884e-03 8.35168809e-02 1.02422405e-02 4.82206140e-03 9.31329876e-02 2.51863953e-02 1.13898702e-03 8.22103098e-02 1.60372816e-02 1.15551911e-05 7.99940433e-03 1.02275699e-01 2.85451766e-03 4.36407104e-02 4.31873445e-07 6.57288954e-02 9.39603371e-04 3.13387588e-02 6.17908197e-04 2.92181066e-04 9.78421345e-02 6.07657023e-02 6.40396178e-02 4.62569783e-12 7.83742033e-03 1.68831123e-03 1.84474970e-06 3.66242370e-04 1.79807048e-05 7.59147247e-03 1.32889207e-02 3.80041972e-02] Category counts: [ 5 1 0 219 371 0 9 39 44 398 47 24 495 115 7 384 104 0 48 499 15 245 0 347 5 152 0 4 490 300 320 0 41 7 0 7 0 32 68 158] Current means: [56.496956 50.22929 nan 43.21151 72.24021 nan 46.89808 48.515617 48.230076 78.34058 49.06494 51.18458 85.53952 56.319077 44.775166 28.373095 46.263325 nan 51.470585 13.493625 51.108376 61.73073 nan 66.48314 55.253593 53.959602 nan 49.19496 21.414316 38.95648 34.11257 nan 46.325092 47.695198 nan 52.575386 nan 50.305588 51.70619 58.51723 ] Posterior variance: [1.9607843e-01 9.0909088e-01 1.0000000e+01 4.5641260e-03 2.6946913e-03 1.0000000e+01 1.0989010e-01 2.5575448e-02 2.2675738e-02 2.5119316e-03 2.1231424e-02 4.1493773e-02 2.0197940e-03 8.6880978e-03 1.4084508e-01 2.6034887e-03 9.6061481e-03 1.0000000e+01 2.0790022e-02 2.0036064e-03 6.6225164e-02 4.0799673e-03 1.0000000e+01 2.8810140e-03 1.9607843e-01 6.5746219e-03 1.0000000e+01 2.4390244e-01 2.0403999e-03 3.3322226e-03 3.1240238e-03 1.0000000e+01 2.4330901e-02 1.4084508e-01 1.0000000e+01 1.4084508e-01 1.0000000e+01 3.1152649e-02 1.4684288e-02 6.3251103e-03] Posterior means: [56.369564 50.208447 nan 43.214607 72.23422 nan 46.932167 48.519413 48.234093 78.33346 49.066925 51.179665 85.53235 56.313587 44.84875 28.378725 46.266914 nan 51.46753 13.500939 51.101036 61.725945 nan 66.478386 55.15058 53.956997 nan 49.2146 21.420149 38.96016 34.117535 nan 46.334034 47.727657 nan 52.539112 nan 50.304634 51.703682 58.51184 ] Resampled means: [56.6996 51.93298 nan 43.211132 72.23603 nan 46.89099 48.520893 48.21903 78.33005 49.09784 51.131424 85.5351 56.329945 44.904953 28.37801 46.26097 nan 51.476604 13.499941 51.23373 61.734695 nan 66.47641 54.963127 53.955772 nan 49.324627 21.419565 38.96012 34.119423 nan 46.35519 47.82891 nan 52.861965 nan 50.265095 51.71132 58.519913] Chosen means: [56.6996 51.93298 50.256966 43.211132 72.23603 50.042118 46.89099 48.520893 48.21903 78.33005 49.09784 51.131424 85.5351 56.329945 44.904953 28.37801 46.26097 52.33351 51.476604 13.499941 51.23373 61.734695 48.401085 66.47641 54.963127 53.955772 50.394512 49.324627 21.419565 38.96012 34.119423 50.328022 46.35519 47.82891 53.050323 52.861965 50.84937 50.265095 51.71132 58.519913] Category counts: [ 3 0 0 232 371 3 7 38 51 389 56 23 490 118 6 384 89 0 38 500 18 248 0 347 7 162 3 1 487 309 313 0 42 8 0 4 0 20 69 164] New alpha: [3.312500e+00 3.125000e-01 3.125000e-01 2.323125e+02 3.713125e+02 3.312500e+00 7.312500e+00 3.831250e+01 5.131250e+01 3.893125e+02 5.631250e+01 2.331250e+01 4.903125e+02 1.183125e+02 6.312500e+00 3.843125e+02 8.931250e+01 3.125000e-01 3.831250e+01 5.003125e+02 1.831250e+01 2.483125e+02 3.125000e-01 3.473125e+02 7.312500e+00 1.623125e+02 3.312500e+00 1.312500e+00 4.873125e+02 3.093125e+02 3.133125e+02 3.125000e-01 4.231250e+01 8.312500e+00 3.125000e-01 4.312500e+00 3.125000e-01 2.031250e+01 6.931250e+01 1.643125e+02] New probs: [3.21530708e-04 7.39561801e-06 2.72797070e-06 4.66700606e-02 7.50734210e-02 7.27361417e-04 1.42301631e-03 7.55701726e-03 9.05699469e-03 7.75841475e-02 1.20473625e-02 3.73163773e-03 9.52834859e-02 2.66124364e-02 8.42647860e-04 8.24900120e-02 1.67048480e-02 2.21223573e-09 6.75999466e-03 1.05529778e-01 4.88654384e-03 5.01871258e-02 2.52224963e-05 6.41018823e-02 1.19888119e-03 3.09709180e-02 5.73549187e-04 5.05295407e-04 9.53043476e-02 5.55820763e-02 6.65273815e-02 1.36461892e-07 1.07151661e-02 1.90410821e-03 1.11817772e-05 8.39246262e-04 1.46143757e-05 4.24549496e-03 1.36724915e-02 3.03088091e-02] Category counts: [ 3 0 0 232 371 3 7 38 51 389 56 23 490 118 6 384 89 0 38 500 18 248 0 347 7 162 3 1 487 309 313 0 42 8 0 4 0 20 69 164] Current means: [56.463425 nan nan 43.297874 72.458046 50.300884 47.326252 48.419506 48.334663 78.516174 49.055294 50.869194 85.57444 56.46686 45.699535 28.34627 46.348454 nan 51.530304 13.500216 50.93369 61.939987 nan 66.67784 55.421215 53.972534 51.78404 48.462624 21.410202 38.883747 34.012566 nan 46.306858 47.868473 nan 52.51217 nan 50.654495 51.716213 58.573505] Posterior variance: [3.22580665e-01 1.00000000e+01 1.00000000e+01 4.30848775e-03 2.69469130e-03 3.22580665e-01 1.40845075e-01 2.62467209e-02 1.95694715e-02 2.57003331e-03 1.78253129e-02 4.32900414e-02 2.04039994e-03 8.46740045e-03 1.63934425e-01 2.60348874e-03 1.12233451e-02 1.00000000e+01 2.62467209e-02 1.99960009e-03 5.52486181e-02 4.03063279e-03 1.00000000e+01 2.88101402e-03 1.40845075e-01 6.16903137e-03 3.22580665e-01 9.09090877e-01 2.05296651e-03 3.23519879e-03 3.19386763e-03 1.00000000e+01 2.37529706e-02 1.23456784e-01 1.00000000e+01 2.43902445e-01 1.00000000e+01 4.97512445e-02 1.44717805e-02 6.09384477e-03] Posterior means: [56.25493 nan nan 43.30076 72.451996 50.29118 47.363907 48.423653 48.33792 78.50884 49.056976 50.865433 85.567184 56.461384 45.77003 28.351908 46.35255 nan 51.526287 13.507514 50.92853 61.935177 nan 66.673035 55.344856 53.970085 51.72649 48.602386 21.416073 38.887344 34.017673 nan 46.315628 47.894787 nan 52.450897 nan 50.651237 51.713726 58.56828 ] Resampled means: [55.69862 nan nan 43.31078 72.45312 50.153187 47.30779 48.41313 48.348423 78.5107 49.073395 50.96239 85.56707 56.455288 45.808617 28.351074 46.34356 nan 51.513474 13.507926 50.92774 61.93856 nan 66.67193 55.451523 53.97859 51.9743 48.747993 21.416214 38.89711 34.01872 nan 46.3242 47.696896 nan 52.15747 nan 50.69268 51.74144 58.575657] Chosen means: [55.69862 51.93298 50.256966 43.31078 72.45312 50.153187 47.30779 48.41313 48.348423 78.5107 49.073395 50.96239 85.56707 56.455288 45.808617 28.351074 46.34356 52.33351 51.513474 13.507926 50.92774 61.93856 48.401085 66.67193 55.451523 53.97859 51.9743 48.747993 21.416214 38.89711 34.01872 50.328022 46.3242 47.696896 53.050323 52.15747 50.84937 50.69268 51.74144 58.575657] Category counts: [ 3 0 0 230 376 4 4 36 41 387 65 24 486 134 5 386 80 0 37 500 26 257 0 341 6 149 2 4 483 310 313 0 51 13 0 3 0 26 66 152] New alpha: [3.312500e+00 3.125000e-01 3.125000e-01 2.303125e+02 3.763125e+02 4.312500e+00 4.312500e+00 3.631250e+01 4.131250e+01 3.873125e+02 6.531250e+01 2.431250e+01 4.863125e+02 1.343125e+02 5.312500e+00 3.863125e+02 8.031250e+01 3.125000e-01 3.731250e+01 5.003125e+02 2.631250e+01 2.573125e+02 3.125000e-01 3.413125e+02 6.312500e+00 1.493125e+02 2.312500e+00 4.312500e+00 4.833125e+02 3.103125e+02 3.133125e+02 3.125000e-01 5.131250e+01 1.331250e+01 3.125000e-01 3.312500e+00 3.125000e-01 2.631250e+01 6.631250e+01 1.523125e+02] New probs: [1.5712998e-04 5.9312595e-05 9.6077425e-04 4.5810871e-02 7.0267960e-02 1.3380859e-03 1.1873130e-03 8.0016395e-03 7.8140050e-03 7.9072870e-02 1.2927977e-02 4.3127136e-03 9.7659752e-02 2.5753029e-02 2.1056326e-03 7.7587143e-02 2.0465082e-02 6.0544262e-06 7.3906374e-03 9.9876143e-02 4.6110875e-03 5.7262525e-02 8.3207913e-11 6.7918070e-02 9.0016040e-04 2.4404518e-02 6.4713042e-04 6.8338146e-04 1.0264046e-01 5.5605642e-02 5.8058236e-02 2.7662711e-05 1.0486388e-02 2.0336760e-03 6.3426887e-05 4.4843627e-04 2.0375089e-06 5.1144604e-03 1.3618460e-02 3.2720357e-02] Category counts: [ 3 0 0 230 376 4 4 36 41 387 65 24 486 134 5 386 80 0 37 500 26 257 0 341 6 149 2 4 483 310 313 0 51 13 0 3 0 26 66 152] Current means: [55.699356 nan nan 43.28614 72.52094 49.366817 46.583244 48.471878 48.344746 78.60422 49.118282 51.061115 85.60139 56.44649 46.077045 28.29696 46.212807 nan 51.432087 13.500216 50.843338 61.999496 nan 66.72551 55.7066 53.97209 50.869637 48.41204 21.381493 38.872723 33.972393 nan 46.027866 48.089928 nan 52.655346 nan 50.942593 51.79045 58.68547 ] Posterior variance: [3.2258067e-01 1.0000000e+01 1.0000000e+01 4.3459362e-03 2.6588673e-03 2.4390244e-01 2.4390244e-01 2.7700832e-02 2.4330901e-02 2.5833119e-03 1.5360983e-02 4.1493773e-02 2.0571898e-03 7.4571213e-03 1.9607843e-01 2.5900025e-03 1.2484395e-02 1.0000000e+01 2.6954180e-02 1.9996001e-03 3.8314175e-02 3.8895370e-03 1.0000000e+01 2.9316915e-03 1.6393442e-01 6.7069079e-03 4.7619051e-01 2.4390244e-01 2.0699648e-03 3.2247661e-03 3.1938676e-03 1.0000000e+01 1.9569471e-02 7.6335877e-02 1.0000000e+01 3.2258067e-01 1.0000000e+01 3.8314175e-02 1.5128593e-02 6.5746219e-03] Posterior means: [55.515507 nan nan 43.28906 72.51496 49.382263 46.66658 48.476112 48.348774 78.59683 49.119637 51.05671 85.594055 56.441685 46.153965 28.302582 46.217537 nan 51.428223 13.507514 50.840103 61.99483 nan 66.720604 55.613045 53.96943 50.828224 48.450775 21.387415 38.87631 33.977512 nan 46.03564 48.104507 nan 52.56969 nan 50.93898 51.78774 58.679756] Resampled means: [55.55832 nan nan 43.29375 72.52015 49.080017 47.034515 48.436176 48.336613 78.599724 49.10584 51.134254 85.59443 56.437504 46.074757 28.301647 46.231533 nan 51.429863 13.509312 50.86279 61.99022 nan 66.71808 55.671417 53.963844 50.187183 48.187138 21.388847 38.870457 33.9773 nan 46.057983 48.108913 nan 52.4439 nan 50.943485 51.781773 58.676426] Chosen means: [55.55832 51.93298 50.256966 43.29375 72.52015 49.080017 47.034515 48.436176 48.336613 78.599724 49.10584 51.134254 85.59443 56.437504 46.074757 28.301647 46.231533 52.33351 51.429863 13.509312 50.86279 61.99022 48.401085 66.71808 55.671417 53.963844 50.187183 48.187138 21.388847 38.870457 33.9773 50.328022 46.057983 48.108913 53.050323 52.4439 50.84937 50.943485 51.781773 58.676426] Category counts: [ 0 1 6 222 373 12 3 37 45 392 49 11 483 149 10 387 96 0 46 500 23 265 0 333 6 139 5 2 479 310 314 0 51 8 0 4 0 18 78 143] New alpha: [3.125000e-01 1.312500e+00 6.312500e+00 2.223125e+02 3.733125e+02 1.231250e+01 3.312500e+00 3.731250e+01 4.531250e+01 3.923125e+02 4.931250e+01 1.131250e+01 4.833125e+02 1.493125e+02 1.031250e+01 3.873125e+02 9.631250e+01 3.125000e-01 4.631250e+01 5.003125e+02 2.331250e+01 2.653125e+02 3.125000e-01 3.333125e+02 6.312500e+00 1.393125e+02 5.312500e+00 2.312500e+00 4.793125e+02 3.103125e+02 3.143125e+02 3.125000e-01 5.131250e+01 8.312500e+00 3.125000e-01 4.312500e+00 3.125000e-01 1.831250e+01 7.831250e+01 1.433125e+02] New probs: [3.2528493e-07 9.7315654e-04 1.1290865e-03 4.4298220e-02 7.0935808e-02 2.1667371e-03 2.0861684e-03 7.7117491e-03 8.4956754e-03 8.3362542e-02 9.8740179e-03 2.7674797e-03 9.6617669e-02 3.1818144e-02 2.1889750e-03 7.6117858e-02 2.1244030e-02 8.5643493e-05 8.4548257e-03 9.9203341e-02 6.0355612e-03 5.6050289e-02 4.2908840e-05 6.5489225e-02 1.0757667e-03 2.7090246e-02 1.3810359e-03 5.6426169e-04 8.9673921e-02 5.9869163e-02 6.5467805e-02 2.6206519e-05 9.9636735e-03 1.9069095e-03 8.2215356e-06 1.1203089e-03 7.8103127e-05 1.6792762e-03 1.5977500e-02 2.6967999e-02] Category counts: [ 0 1 6 222 373 12 3 37 45 392 49 11 483 149 10 387 96 0 46 500 23 265 0 333 6 139 5 2 479 310 314 0 51 8 0 4 0 18 78 143] Current means: [ nan 52.479576 50.20181 43.18207 72.506905 48.715897 47.03783 48.443245 48.407345 78.61407 49.269127 50.725163 85.62419 56.47334 45.87961 28.233189 46.342896 nan 51.72117 13.500216 50.974247 62.095158 nan 66.79538 55.7232 54.03293 50.496967 48.19935 21.354717 38.834824 33.938824 nan 46.053265 47.856354 nan 51.94744 nan 50.96286 51.82108 58.7547 ] Posterior variance: [1.00000000e+01 9.09090877e-01 1.63934425e-01 4.50247619e-03 2.68024649e-03 8.26446265e-02 3.22580665e-01 2.69541796e-02 2.21729502e-02 2.55036983e-03 2.03665998e-02 9.00900885e-02 2.06996477e-03 6.70690788e-03 9.90098938e-02 2.58331187e-03 1.04058273e-02 1.00000000e+01 2.16919743e-02 1.99960009e-03 4.32900414e-02 3.77216144e-03 1.00000000e+01 3.00210132e-03 1.63934425e-01 7.18907220e-03 1.96078435e-01 4.76190507e-01 2.08724686e-03 3.22476611e-03 3.18369945e-03 1.00000000e+01 1.95694715e-02 1.23456784e-01 1.00000000e+01 2.43902445e-01 1.00000000e+01 5.52486181e-02 1.28040975e-02 6.98811980e-03] Posterior means: [ nan 52.25416 50.198498 43.185143 72.50088 48.726513 47.133385 48.447437 48.410873 78.606766 49.270615 50.71863 85.61682 56.468998 45.9204 28.238813 46.3467 nan 51.717434 13.507514 50.97003 62.09059 nan 66.79033 55.629375 54.030025 50.487225 48.28509 21.360697 38.838425 33.94394 nan 46.06099 47.882816 nan 51.899944 nan 50.957542 51.818752 58.748577] Resampled means: [ nan 50.96344 49.924606 43.18251 72.50734 48.856636 46.910984 48.50051 48.411167 78.60345 49.269382 50.683773 85.61384 56.465603 45.987354 28.237478 46.335476 nan 51.74293 13.505055 50.95849 62.09311 nan 66.78845 55.56518 54.036335 50.64767 48.359955 21.35861 38.83482 33.94237 nan 46.04602 47.848473 nan 51.892124 nan 50.928 51.831867 58.736877] Chosen means: [55.55832 50.96344 49.924606 43.18251 72.50734 48.856636 46.910984 48.50051 48.411167 78.60345 49.269382 50.683773 85.61384 56.465603 45.987354 28.237478 46.335476 52.33351 51.74293 13.505055 50.95849 62.09311 48.401085 66.78845 55.56518 54.036335 50.64767 48.359955 21.35861 38.83482 33.94237 50.328022 46.04602 47.848473 53.050323 51.892124 50.84937 50.928 51.831867 58.736877] Category counts: [ 0 4 3 228 374 9 10 28 49 399 36 11 476 151 12 396 102 0 32 501 43 265 0 323 6 141 5 3 472 314 304 0 45 6 0 6 0 6 89 151] New alpha: [3.125000e-01 4.312500e+00 3.312500e+00 2.283125e+02 3.743125e+02 9.312500e+00 1.031250e+01 2.831250e+01 4.931250e+01 3.993125e+02 3.631250e+01 1.131250e+01 4.763125e+02 1.513125e+02 1.231250e+01 3.963125e+02 1.023125e+02 3.125000e-01 3.231250e+01 5.013125e+02 4.331250e+01 2.653125e+02 3.125000e-01 3.233125e+02 6.312500e+00 1.413125e+02 5.312500e+00 3.312500e+00 4.723125e+02 3.143125e+02 3.043125e+02 3.125000e-01 4.531250e+01 6.312500e+00 3.125000e-01 6.312500e+00 3.125000e-01 6.312500e+00 8.931250e+01 1.513125e+02] New probs: [2.57881300e-10 5.03240270e-04 3.31567950e-04 4.79556844e-02 7.11394921e-02 1.65699259e-03 2.45738286e-03 7.80669786e-03 1.02805523e-02 7.49592036e-02 7.58781005e-03 1.95630430e-03 8.94299224e-02 3.39032486e-02 3.19130626e-03 8.29987526e-02 2.06685662e-02 4.83961600e-08 7.93622900e-03 1.06288001e-01 9.49819572e-03 5.31287715e-02 1.02058828e-09 6.87851533e-02 1.43412524e-03 2.48288196e-02 7.27112056e-04 3.23668879e-04 9.46266428e-02 5.98968044e-02 6.07028008e-02 1.54615498e-09 9.25386045e-03 1.26554887e-03 1.44748365e-05 1.10961159e-03 5.02293233e-07 6.97192270e-04 1.55954864e-02 2.70604603e-02] Category counts: [ 0 4 3 228 374 9 10 28 49 399 36 11 476 151 12 396 102 0 32 501 43 265 0 323 6 141 5 3 472 314 304 0 45 6 0 6 0 6 89 151] Current means: [ nan 50.956047 49.70046 43.172142 72.501564 48.88909 46.924988 48.792908 48.37741 78.681404 49.56804 50.47124 85.6736 56.462135 45.612247 28.205046 46.32508 nan 51.70361 13.509375 50.49199 62.22778 nan 66.83861 56.089733 53.993717 50.631584 48.47998 21.31609 38.769188 33.901714 nan 46.27619 48.01192 nan 51.830814 nan 51.23631 51.79422 58.901142] Posterior variance: [1.0000000e+01 2.4390244e-01 3.2258067e-01 4.3840418e-03 2.6730821e-03 1.0989010e-01 9.9009894e-02 3.5587188e-02 2.0366600e-02 2.5056377e-03 2.7700832e-02 9.0090089e-02 2.1003990e-03 6.6181333e-03 8.2644626e-02 2.5246150e-03 9.7943190e-03 1.0000000e+01 3.1152649e-02 1.9956096e-03 2.3201857e-02 3.7721614e-03 1.0000000e+01 3.0950170e-03 1.6393442e-01 7.0871720e-03 1.9607843e-01 3.2258067e-01 2.1181952e-03 3.1836994e-03 3.2883920e-03 1.0000000e+01 2.2172950e-02 1.6393442e-01 1.0000000e+01 1.6393442e-01 1.0000000e+01 1.6393442e-01 1.1223345e-02 6.6181333e-03] Posterior means: [ nan 50.93273 49.71012 43.175137 72.49555 48.901302 46.95543 48.797207 48.380714 78.67422 49.569237 50.466995 85.66611 56.457863 45.64851 28.210546 46.328682 nan 51.698303 13.516657 50.490852 62.223164 nan 66.8334 55.9899 53.990883 50.6192 48.529015 21.322166 38.772766 33.90701 nan 46.284447 48.04451 nan 51.8008 nan 51.216038 51.792206 58.895256] Resampled means: [ nan 50.874355 49.63321 43.170856 72.497116 48.859585 46.91243 48.8109 48.366768 78.673065 49.519245 50.548435 85.66423 56.448452 45.701645 28.212738 46.32994 nan 51.690792 13.516532 50.507423 62.221493 nan 66.8344 55.620174 53.987232 50.653397 48.583878 21.321556 38.773106 33.90619 nan 46.273636 47.960907 nan 51.72347 nan 51.13991 51.80767 58.892677] Chosen means: [55.55832 50.874355 49.63321 43.170856 72.497116 48.859585 46.91243 48.8109 48.366768 78.673065 49.519245 50.548435 85.66423 56.448452 45.701645 28.212738 46.32994 52.33351 51.690792 13.516532 50.507423 62.221493 48.401085 66.8344 55.620174 53.987232 50.653397 48.583878 21.321556 38.773106 33.90619 50.328022 46.273636 47.960907 53.050323 51.72347 50.84937 51.13991 51.80767 58.892677] Category counts: [ 0 3 2 232 372 12 12 41 51 405 37 5 469 149 16 387 106 0 41 499 47 264 0 317 6 145 5 2 478 314 301 0 27 5 0 9 0 2 76 163] New alpha: [3.125000e-01 3.312500e+00 2.312500e+00 2.323125e+02 3.723125e+02 1.231250e+01 1.231250e+01 4.131250e+01 5.131250e+01 4.053125e+02 3.731250e+01 5.312500e+00 4.693125e+02 1.493125e+02 1.631250e+01 3.873125e+02 1.063125e+02 3.125000e-01 4.131250e+01 4.993125e+02 4.731250e+01 2.643125e+02 3.125000e-01 3.173125e+02 6.312500e+00 1.453125e+02 5.312500e+00 2.312500e+00 4.783125e+02 3.143125e+02 3.013125e+02 3.125000e-01 2.731250e+01 5.312500e+00 3.125000e-01 9.312500e+00 3.125000e-01 2.312500e+00 7.631250e+01 1.633125e+02] New probs: [8.1305305e-05 4.0257012e-04 5.0905894e-04 4.7141381e-02 7.2554350e-02 2.4364113e-03 2.5781791e-03 8.3703976e-03 1.1224793e-02 7.7730633e-02 6.5033520e-03 1.3970666e-03 8.5700773e-02 3.0159622e-02 3.4838710e-03 8.0795169e-02 2.2048309e-02 1.6325093e-05 1.1585725e-02 9.9806443e-02 8.8787004e-03 5.1681787e-02 3.3559558e-05 6.4786375e-02 1.3162036e-03 3.1689201e-02 1.1220352e-03 2.2890967e-04 9.4658867e-02 5.9853178e-02 6.2190644e-02 5.2058695e-06 7.9280436e-03 1.6935377e-03 4.0660449e-04 2.6941481e-03 1.3713109e-04 7.9899881e-04 1.2348598e-02 3.3023015e-02] Category counts: [ 0 3 2 232 372 12 12 41 51 405 37 5 469 149 16 387 106 0 41 499 47 264 0 317 6 145 5 2 478 314 301 0 27 5 0 9 0 2 76 163] Current means: [ nan 51.911312 50.41223 43.05192 72.52873 48.137142 47.229504 48.536236 48.356457 78.752464 49.551575 49.539135 85.722 56.44187 45.265587 28.204626 46.276745 nan 51.797386 13.493208 50.562855 62.36721 nan 66.952324 55.918896 53.894665 50.594368 49.080856 21.330736 38.6635 33.78072 nan 46.168358 47.666508 nan 51.540085 nan 50.472237 51.8423 58.997887] Posterior variance: [1.0000000e+01 3.2258067e-01 4.7619051e-01 4.3084878e-03 2.6874496e-03 8.2644626e-02 8.2644626e-02 2.4330901e-02 1.9569471e-02 2.4685264e-03 2.6954180e-02 1.9607843e-01 2.1317415e-03 6.7069079e-03 6.2111799e-02 2.5833119e-03 9.4250711e-03 1.0000000e+01 2.4330901e-02 2.0036064e-03 2.1231424e-02 3.7864444e-03 1.0000000e+01 3.1535793e-03 1.6393442e-01 6.8917987e-03 1.9607843e-01 4.7619051e-01 2.0916127e-03 3.1836994e-03 3.3211557e-03 1.0000000e+01 3.6900368e-02 1.9607843e-01 1.0000000e+01 1.0989010e-01 1.0000000e+01 4.7619051e-01 1.3140605e-02 6.1312076e-03] Posterior means: [ nan 51.84966 50.3926 43.054913 72.522675 48.15254 47.252403 48.5398 48.359673 78.74537 49.55278 49.548172 85.71438 56.437553 45.294994 28.210258 46.28026 nan 51.793015 13.500524 50.56166 62.362526 nan 66.946976 55.82186 53.89198 50.582714 49.124622 21.336731 38.66711 33.786102 nan 46.1825 47.71226 nan 51.523163 nan 50.44975 51.83988 58.99237 ] Resampled means: [ nan 51.534744 50.367676 43.063503 72.5228 48.17123 47.096645 48.48462 48.3405 78.745384 49.56497 49.211876 85.71338 56.42125 45.357674 28.214544 46.270405 nan 51.805775 13.501053 50.548805 62.356987 nan 66.948044 55.68625 53.89256 50.579517 48.546745 21.338078 38.665047 33.78727 nan 46.14147 47.862362 nan 51.391838 nan 50.849964 51.837383 58.992363] Chosen means: [55.55832 51.534744 50.367676 43.063503 72.5228 48.17123 47.096645 48.48462 48.3405 78.745384 49.56497 49.211876 85.71338 56.42125 45.357674 28.214544 46.270405 52.33351 51.805775 13.501053 50.548805 62.356987 48.401085 66.948044 55.68625 53.89256 50.579517 48.546745 21.338078 38.665047 33.78727 50.328022 46.14147 47.862362 53.050323 51.391838 50.84937 50.849964 51.837383 58.992363] Category counts: [ 1 4 1 221 371 13 12 46 43 420 28 11 454 138 22 391 98 0 51 500 49 266 0 304 8 138 6 3 476 327 286 0 37 8 4 8 1 4 64 186] New alpha: [1.312500e+00 4.312500e+00 1.312500e+00 2.213125e+02 3.713125e+02 1.331250e+01 1.231250e+01 4.631250e+01 4.331250e+01 4.203125e+02 2.831250e+01 1.131250e+01 4.543125e+02 1.383125e+02 2.231250e+01 3.913125e+02 9.831250e+01 3.125000e-01 5.131250e+01 5.003125e+02 4.931250e+01 2.663125e+02 3.125000e-01 3.043125e+02 8.312500e+00 1.383125e+02 6.312500e+00 3.312500e+00 4.763125e+02 3.273125e+02 2.863125e+02 3.125000e-01 3.731250e+01 8.312500e+00 4.312500e+00 8.312500e+00 1.312500e+00 4.312500e+00 6.431250e+01 1.863125e+02] New probs: [3.1579746e-04 1.2034466e-03 4.1637267e-04 4.5945607e-02 7.4523404e-02 2.8895517e-03 2.2052177e-03 9.3941744e-03 1.0714262e-02 8.2592189e-02 7.8994324e-03 1.5829586e-03 8.4145099e-02 2.9569726e-02 3.4492167e-03 7.5718820e-02 1.9744534e-02 2.7605086e-08 9.7978413e-03 9.6142828e-02 1.2056282e-02 5.2694034e-02 1.9718686e-04 5.7571933e-02 2.1654437e-03 3.0535659e-02 7.9975463e-04 7.8606291e-04 9.9203959e-02 6.2102456e-02 6.0355555e-02 4.7684560e-05 9.5267855e-03 1.7416665e-03 5.7588954e-04 1.1824957e-03 2.2736624e-04 2.4070423e-04 1.4205249e-02 3.5533451e-02]
Category counts: [ 1 4 1 221 371 13 12 46 43 420 28 11 454 138 22 391 98 0 51 500 49 266 0 304 8 138 6 3 476 327 286 0 37 8 4 8 1 4 64 186] Current means: [54.09279 51.61155 48.970226 43.005524 72.53725 47.893913 46.940582 48.537926 48.31994 78.87615 49.646786 49.617313 85.83809 56.270332 45.264362 28.214853 46.154694 nan 51.607475 13.502434 50.60423 62.569286 nan 67.071045 55.979755 53.99313 50.358555 49.330902 21.332085 38.587025 33.70303 nan 46.171238 47.64765 53.32832 51.371994 52.137756 50.968536 51.909203 59.000492] Posterior variance: [9.09090877e-01 2.43902445e-01 9.09090877e-01 4.52284003e-03 2.69469130e-03 7.63358772e-02 8.26446265e-02 2.16919743e-02 2.32018568e-02 2.38038553e-03 3.55871879e-02 9.00900885e-02 2.20215810e-03 7.24112941e-03 4.52488698e-02 2.55689071e-03 1.01936804e-02 1.00000000e+01 1.95694715e-02 1.99960009e-03 2.03665998e-02 3.75798554e-03 1.00000000e+01 3.28839198e-03 1.23456784e-01 7.24112941e-03 1.63934425e-01 3.22580665e-01 2.10039900e-03 3.05716903e-03 3.49528133e-03 1.00000000e+01 2.69541796e-02 1.23456784e-01 2.43902445e-01 1.23456784e-01 9.09090877e-01 2.43902445e-01 1.56006245e-02 5.37345512e-03] Posterior means: [53.72072 51.572247 49.063847 43.008682 72.53118 47.90999 46.96587 48.541096 48.32384 78.86928 49.648045 49.62076 85.83019 56.265793 45.28579 28.220423 46.15861 nan 51.60433 13.509732 50.602997 62.564564 nan 67.06543 55.90593 53.99024 50.352673 49.35249 21.338106 38.59051 33.708725 nan 46.181557 47.676693 53.247143 51.355057 51.943417 50.944916 51.906223 58.99565 ] Resampled means: [54.14435 51.51574 50.225204 43.007874 72.53262 47.823124 46.886475 48.541153 48.330833 78.868614 49.63132 49.610664 85.82778 56.26926 45.290638 28.221455 46.163025 nan 51.597992 13.510991 50.632824 62.560715 nan 67.06664 55.741848 53.995968 50.07911 49.990967 21.337154 38.590508 33.710644 nan 46.171986 47.92779 53.10456 51.400986 51.01554 51.334644 51.925224 58.998425] Chosen means: [54.14435 51.51574 50.225204 43.007874 72.53262 47.823124 46.886475 48.541153 48.330833 78.868614 49.63132 49.610664 85.82778 56.26926 45.290638 28.221455 46.163025 52.33351 51.597992 13.510991 50.632824 62.560715 48.401085 67.06664 55.741848 53.995968 50.07911 49.990967 21.337154 38.590508 33.710644 50.328022 46.171986 47.92779 53.10456 51.400986 51.01554 51.334644 51.925224 58.998425] Category counts: [ 1 4 0 231 372 11 6 46 58 422 35 9 450 145 13 389 97 0 52 500 50 271 0 285 9 138 2 4 474 322 291 0 40 8 2 3 2 2 64 192] New alpha: [1.312500e+00 4.312500e+00 3.125000e-01 2.313125e+02 3.723125e+02 1.131250e+01 6.312500e+00 4.631250e+01 5.831250e+01 4.223125e+02 3.531250e+01 9.312500e+00 4.503125e+02 1.453125e+02 1.331250e+01 3.893125e+02 9.731250e+01 3.125000e-01 5.231250e+01 5.003125e+02 5.031250e+01 2.713125e+02 3.125000e-01 2.853125e+02 9.312500e+00 1.383125e+02 2.312500e+00 4.312500e+00 4.743125e+02 3.223125e+02 2.913125e+02 3.125000e-01 4.031250e+01 8.312500e+00 2.312500e+00 3.312500e+00 2.312500e+00 2.312500e+00 6.431250e+01 1.923125e+02] New probs: [4.2126368e-05 9.8625361e-04 1.2499719e-05 4.1902453e-02 8.0743097e-02 1.8524257e-03 5.6517887e-04 1.0637982e-02 1.1753799e-02 8.0684826e-02 5.9381477e-03 2.0840615e-03 9.5920287e-02 3.1153545e-02 2.5282053e-03 7.2370842e-02 2.0604830e-02 5.5813422e-07 9.7625107e-03 9.6488796e-02 1.0915829e-02 5.4293714e-02 2.1782221e-06 5.9997138e-02 1.5722847e-03 2.7767288e-02 1.6390810e-04 6.8291789e-04 9.1297016e-02 6.2877029e-02 5.7574734e-02 4.6575416e-05 8.9534540e-03 1.5097510e-03 7.5386581e-04 1.6484876e-03 2.8188503e-04 6.3175359e-04 1.3671605e-02 3.9325885e-02] Category counts: [ 1 4 0 231 372 11 6 46 58 422 35 9 450 145 13 389 97 0 52 500 50 271 0 285 9 138 2 4 474 322 291 0 40 8 2 3 2 2 64 192] Current means: [54.02542 51.64749 nan 42.99082 72.563896 47.444324 46.767925 48.67489 48.369507 78.9213 49.408688 49.616497 85.86962 56.292336 45.3683 28.168617 46.284916 nan 51.509068 13.500216 50.734047 62.833134 nan 67.251465 55.32457 54.090794 50.638016 50.518047 21.31873 38.559723 33.683533 nan 45.860584 48.328003 53.633514 51.70387 50.363575 50.463818 52.08528 59.235855] Posterior variance: [9.09090877e-01 2.43902445e-01 1.00000000e+01 4.32713097e-03 2.68744957e-03 9.00900885e-02 1.63934425e-01 2.16919743e-02 1.72117036e-02 2.36910675e-03 2.84900293e-02 1.09890103e-01 2.22172844e-03 6.89179869e-03 7.63358772e-02 2.57003331e-03 1.02986610e-02 1.00000000e+01 1.91938579e-02 1.99960009e-03 1.99600812e-02 3.68867558e-03 1.00000000e+01 3.50754103e-03 1.09890103e-01 7.24112941e-03 4.76190507e-01 2.43902445e-01 2.10925960e-03 3.10462574e-03 3.43524548e-03 1.00000000e+01 2.49376576e-02 1.23456784e-01 4.76190507e-01 3.22580665e-01 4.76190507e-01 4.76190507e-01 1.56006245e-02 5.20562194e-03] Posterior means: [53.659477 51.60731 nan 42.99386 72.55783 47.46735 46.820908 48.67776 48.372314 78.91445 49.41037 49.620712 85.86166 56.288 45.403656 28.174227 46.288746 nan 51.506172 13.507514 50.73258 62.828403 nan 67.24542 55.26606 54.087833 50.60763 50.505413 21.32478 38.563274 33.68914 nan 45.870907 48.348644 53.460487 51.648907 50.34626 50.44173 52.082027 59.23105 ] Resampled means: [54.406803 51.028355 nan 42.99867 72.55519 47.427605 46.910137 48.67518 48.363815 78.91607 49.427994 49.508144 85.85906 56.28747 45.40935 28.169222 46.282005 nan 51.509872 13.506594 50.743168 62.833218 nan 67.24621 55.364838 54.091442 50.6718 50.62326 21.327293 38.56305 33.687893 nan 45.873043 48.326405 53.314575 52.164066 50.931313 50.03181 52.074623 59.226696] Chosen means: [54.406803 51.028355 50.225204 42.99867 72.55519 47.427605 46.910137 48.67518 48.363815 78.91607 49.427994 49.508144 85.85906 56.28747 45.40935 28.169222 46.282005 52.33351 51.509872 13.506594 50.743168 62.833218 48.401085 67.24621 55.364838 54.091442 50.6718 50.62326 21.327293 38.56305 33.687893 50.328022 45.873043 48.326405 53.314575 52.164066 50.931313 50.03181 52.074623 59.226696] Category counts: [ 0 3 0 216 373 6 6 50 55 427 29 9 447 150 8 381 117 0 54 498 50 243 0 280 10 119 2 4 476 316 299 0 48 6 4 11 0 1 81 221] New alpha: [3.125000e-01 3.312500e+00 3.125000e-01 2.163125e+02 3.733125e+02 6.312500e+00 6.312500e+00 5.031250e+01 5.531250e+01 4.273125e+02 2.931250e+01 9.312500e+00 4.473125e+02 1.503125e+02 8.312500e+00 3.813125e+02 1.173125e+02 3.125000e-01 5.431250e+01 4.983125e+02 5.031250e+01 2.433125e+02 3.125000e-01 2.803125e+02 1.031250e+01 1.193125e+02 2.312500e+00 4.312500e+00 4.763125e+02 3.163125e+02 2.993125e+02 3.125000e-01 4.831250e+01 6.312500e+00 4.312500e+00 1.131250e+01 3.125000e-01 1.312500e+00 8.131250e+01 2.213125e+02] New probs: [1.0027899e-05 9.6206251e-04 3.8609134e-05 3.9303672e-02 7.3366627e-02 9.9128485e-04 1.9634638e-03 7.7797603e-03 1.1180174e-02 8.9896031e-02 6.8137515e-03 1.4427023e-03 8.6484656e-02 3.1512912e-02 1.2425840e-03 7.6020107e-02 2.2720758e-02 2.1456301e-06 1.1375427e-02 9.8592363e-02 1.2327971e-02 5.0267212e-02 1.3819245e-05 5.7189777e-02 1.5199475e-03 2.2993298e-02 3.4571928e-04 3.5937462e-04 9.2499122e-02 6.0978580e-02 6.6228621e-02 1.6300258e-04 8.7771183e-03 1.2976262e-03 9.9095120e-04 2.9385914e-03 4.6898512e-07 1.0067152e-04 1.5397001e-02 4.3912262e-02]
Out[5]:
Plotting results
In [6]:
Copied!
# Prepare data for the animation
data_points = datapoints["datapoints", "obs"].tolist()
np.random.seed(42)
jitter = np.random.uniform(-0.05, 0.05, size=len(data_points)).tolist()
std_dev = np.sqrt(OBS_VARIANCE) * 1.5
all_cluster_assignments_list = [a.tolist() for a in all_cluster_assignment]
all_posterior_means_list = [m.tolist() for m in all_posterior_means]
all_posterior_weights_list = [w.tolist() for w in all_posterior_weights]
# Define a consistent color palette to use throughout the visualization
color_palette = """
const plotColors = [
"#4c78a8", "#f58518", "#e45756", "#72b7b2", "#54a24b",
"#eeca3b", "#b279a2", "#ff9da6", "#9d755d", "#bab0ac",
"#1f77b4", "#ff7f0e", "#2ca02c", "#d62728", "#9467bd",
"#8c564b", "#e377c2", "#7f7f7f", "#bcbd22", "#17becf"
];
"""
# Shared data initialization for all plot components
frame_data_js = (
"""
// Get current frame data
const frame = $state.frame;
const hoveredCluster = $state.hoveredCluster;
const means = """
+ str(all_posterior_means_list)
+ """[frame];
const weights = """
+ str(all_posterior_weights_list)
+ """[frame];
const assignments = """
+ str(all_cluster_assignments_list)
+ """[frame];
const stdDev = """
+ str(std_dev)
+ """;
"""
)
# Create a visualizer with animation
(
Plot.initialState({"frame": 0, "hoveredCluster": None})
|
# Main visualization that updates based on the current frame
Plot.plot({
"marks": [
# 1. Data points with jitter - show all data points with optional highlighting
Plot.dot(
Plot.js(
"""function() {
"""
+ frame_data_js
+ """
const dataPoints = """
+ str(data_points)
+ """;
const jitter = """
+ str(jitter)
+ """;
"""
+ color_palette
+ """
// Return all points with hover-aware opacity
return dataPoints.map((x, i) => {
const clusterIdx = assignments[i];
// If a cluster is hovered, reduce opacity of other clusters' points
const isHovered = hoveredCluster !== null && clusterIdx === hoveredCluster;
const opacity = hoveredCluster === null ? 0.5 : (isHovered ? 0.7 : 0.15);
return {
x: x,
y: jitter[i],
color: plotColors[clusterIdx % 20],
opacity: opacity
};
});
}()"""
),
{"x": "x", "y": "y", "fill": "color", "r": 3, "opacity": "opacity"},
),
# 2. Combined error bars (both horizontal lines and vertical caps)
Plot.line(
Plot.js(
"""function() {
"""
+ frame_data_js
+ """
const capSize = 0.04; // Size of the vertical cap lines
"""
+ color_palette
+ """
// We'll collect all line segments in a flat array
const result = [];
for (let i = 0; i < means.length; i++) {
// Only include error bars for clusters with weight >= 0.01
if (weights[i] >= 0.01) {
// Determine if this cluster is being hovered
const isHovered = hoveredCluster === i;
const opacity = hoveredCluster === null ? 0.7 : (isHovered ? 1.0 : 0.3);
const strokeWidth = isHovered ? 4 : 3;
const color = plotColors[i % 20];
// Add horizontal line (error bar itself)
result.push({x: means[i] - stdDev, y: 0, cluster: i, color, opacity, width: strokeWidth});
result.push({x: means[i] + stdDev, y: 0, cluster: i, color, opacity, width: strokeWidth});
// Add left cap (vertical line)
result.push({x: means[i] - stdDev, y: -capSize, cluster: i, color, opacity, width: strokeWidth});
result.push({x: means[i] - stdDev, y: capSize, cluster: i, color, opacity, width: strokeWidth});
// Add right cap (vertical line)
result.push({x: means[i] + stdDev, y: -capSize, cluster: i, color, opacity, width: strokeWidth});
result.push({x: means[i] + stdDev, y: capSize, cluster: i, color, opacity, width: strokeWidth});
}
}
return result;
}()"""
),
{
"x": "x",
"y": "y",
"stroke": "color",
"strokeWidth": "width",
"opacity": "opacity",
"z": "cluster",
},
),
# 3. Cluster means as stars
Plot.dot(
Plot.js(
"""function() {
"""
+ frame_data_js
+ """
"""
+ color_palette
+ """
// Create a simple array for each cluster mean
return means.map((mean, i) => {
// Only include means for clusters with sufficient weight
if (weights[i] >= 0.01) {
const isHovered = hoveredCluster === i;
return {
x: mean,
y: 0,
cluster: i,
color: plotColors[i % 20],
opacity: isHovered ? 1.0 : 0.8
};
}
return null; // Skip low-weight clusters
}).filter(d => d !== null); // Remove null values
}()"""
),
{
"x": "x",
"y": "y",
"fill": "color",
"r": 10,
"symbol": "star",
"stroke": "black",
"strokeWidth": 2,
"opacity": "opacity",
},
),
],
"grid": True,
"marginTop": 40,
"marginRight": 40,
"marginBottom": 40,
"marginLeft": 40,
"style": {"height": "400px"},
"title": Plot.js(
"`Dirichlet Mixture Model - Iteration ${$state.frame} of "
+ str(len(all_posterior_means) - 1)
+ "`"
),
"subtitle": "Cluster centers (★) with standard deviation (—) and data points (•)",
})
|
# Animation controls and legend with hover effects
Plot.html([
"div",
{"className": "p-4"},
[
"div",
{"className": "mb-4"},
Plot.Slider(
"frame",
init=0,
range=[0, len(all_posterior_means) - 1],
step=1,
label="Iteration",
width="100%",
fps=8,
),
],
[
"div",
{"className": "mt-4"},
Plot.js(
"""function() {
"""
+ frame_data_js
+ """
// Count assignments in current frame
const counts = {};
assignments.forEach(a => { counts[a] = (counts[a] || 0) + 1; });
"""
+ color_palette
+ """
// Sort clusters by weight, filter by minimum weight, and limit to top 10
const topClusters = Object.keys(weights)
.map(i => ({
id: parseInt(i),
weight: weights[i],
count: counts[parseInt(i)] || 0
}))
.filter(c => c.weight >= 0.01)
.sort((a, b) => b.weight - a.weight)
.slice(0, 10);
// Create placeholder rows for consistent height
const placeholders = Array(Math.max(0, 10 - topClusters.length))
.fill(0)
.map(() => ["tr", {"className": "h-8"}, ["td", {"colSpan": 3}, ""]]);
return [
"div", {},
["h3", {}, `Top Clusters by Weight (Iteration ${frame})`],
["div", {"style": {"height": "280px", "overflow": "auto"}},
["table", {"className": "w-full mt-2"},
["thead", ["tr",
["th", {"className": "text-left"}, "Cluster"],
["th", {"className": "text-left"}, "Weight"],
["th", {"className": "text-left"}, "Points"]
]],
["tbody",
...topClusters.map(cluster =>
["tr", {
"className": "h-8",
"style": {
"cursor": "pointer",
"backgroundColor": $state.hoveredCluster === cluster.id ? "#f0f0f0" : "transparent"
},
"onMouseEnter": () => { $state.hoveredCluster = cluster.id; },
"onMouseLeave": () => { $state.hoveredCluster = null; }
},
["td", {"className": "py-1"},
["div", {"className": "flex items-center"},
["div", {
"style": {
"backgroundColor": plotColors[cluster.id % 20],
"width": "24px",
"height": "24px",
"borderRadius": "4px",
"border": "1px solid rgba(0,0,0,0.2)",
"display": "inline-block",
"marginRight": "8px"
}
}],
`Cluster ${cluster.id}`
]
],
["td", {"className": "py-1"}, cluster.weight.toFixed(4)],
["td", {"className": "py-1"}, cluster.count]
]
),
...placeholders
]
]
]
];
}()"""
),
],
])
)
# Prepare data for the animation
data_points = datapoints["datapoints", "obs"].tolist()
np.random.seed(42)
jitter = np.random.uniform(-0.05, 0.05, size=len(data_points)).tolist()
std_dev = np.sqrt(OBS_VARIANCE) * 1.5
all_cluster_assignments_list = [a.tolist() for a in all_cluster_assignment]
all_posterior_means_list = [m.tolist() for m in all_posterior_means]
all_posterior_weights_list = [w.tolist() for w in all_posterior_weights]
# Define a consistent color palette to use throughout the visualization
color_palette = """
const plotColors = [
"#4c78a8", "#f58518", "#e45756", "#72b7b2", "#54a24b",
"#eeca3b", "#b279a2", "#ff9da6", "#9d755d", "#bab0ac",
"#1f77b4", "#ff7f0e", "#2ca02c", "#d62728", "#9467bd",
"#8c564b", "#e377c2", "#7f7f7f", "#bcbd22", "#17becf"
];
"""
# Shared data initialization for all plot components
frame_data_js = (
"""
// Get current frame data
const frame = $state.frame;
const hoveredCluster = $state.hoveredCluster;
const means = """
+ str(all_posterior_means_list)
+ """[frame];
const weights = """
+ str(all_posterior_weights_list)
+ """[frame];
const assignments = """
+ str(all_cluster_assignments_list)
+ """[frame];
const stdDev = """
+ str(std_dev)
+ """;
"""
)
# Create a visualizer with animation
(
Plot.initialState({"frame": 0, "hoveredCluster": None})
|
# Main visualization that updates based on the current frame
Plot.plot({
"marks": [
# 1. Data points with jitter - show all data points with optional highlighting
Plot.dot(
Plot.js(
"""function() {
"""
+ frame_data_js
+ """
const dataPoints = """
+ str(data_points)
+ """;
const jitter = """
+ str(jitter)
+ """;
"""
+ color_palette
+ """
// Return all points with hover-aware opacity
return dataPoints.map((x, i) => {
const clusterIdx = assignments[i];
// If a cluster is hovered, reduce opacity of other clusters' points
const isHovered = hoveredCluster !== null && clusterIdx === hoveredCluster;
const opacity = hoveredCluster === null ? 0.5 : (isHovered ? 0.7 : 0.15);
return {
x: x,
y: jitter[i],
color: plotColors[clusterIdx % 20],
opacity: opacity
};
});
}()"""
),
{"x": "x", "y": "y", "fill": "color", "r": 3, "opacity": "opacity"},
),
# 2. Combined error bars (both horizontal lines and vertical caps)
Plot.line(
Plot.js(
"""function() {
"""
+ frame_data_js
+ """
const capSize = 0.04; // Size of the vertical cap lines
"""
+ color_palette
+ """
// We'll collect all line segments in a flat array
const result = [];
for (let i = 0; i < means.length; i++) {
// Only include error bars for clusters with weight >= 0.01
if (weights[i] >= 0.01) {
// Determine if this cluster is being hovered
const isHovered = hoveredCluster === i;
const opacity = hoveredCluster === null ? 0.7 : (isHovered ? 1.0 : 0.3);
const strokeWidth = isHovered ? 4 : 3;
const color = plotColors[i % 20];
// Add horizontal line (error bar itself)
result.push({x: means[i] - stdDev, y: 0, cluster: i, color, opacity, width: strokeWidth});
result.push({x: means[i] + stdDev, y: 0, cluster: i, color, opacity, width: strokeWidth});
// Add left cap (vertical line)
result.push({x: means[i] - stdDev, y: -capSize, cluster: i, color, opacity, width: strokeWidth});
result.push({x: means[i] - stdDev, y: capSize, cluster: i, color, opacity, width: strokeWidth});
// Add right cap (vertical line)
result.push({x: means[i] + stdDev, y: -capSize, cluster: i, color, opacity, width: strokeWidth});
result.push({x: means[i] + stdDev, y: capSize, cluster: i, color, opacity, width: strokeWidth});
}
}
return result;
}()"""
),
{
"x": "x",
"y": "y",
"stroke": "color",
"strokeWidth": "width",
"opacity": "opacity",
"z": "cluster",
},
),
# 3. Cluster means as stars
Plot.dot(
Plot.js(
"""function() {
"""
+ frame_data_js
+ """
"""
+ color_palette
+ """
// Create a simple array for each cluster mean
return means.map((mean, i) => {
// Only include means for clusters with sufficient weight
if (weights[i] >= 0.01) {
const isHovered = hoveredCluster === i;
return {
x: mean,
y: 0,
cluster: i,
color: plotColors[i % 20],
opacity: isHovered ? 1.0 : 0.8
};
}
return null; // Skip low-weight clusters
}).filter(d => d !== null); // Remove null values
}()"""
),
{
"x": "x",
"y": "y",
"fill": "color",
"r": 10,
"symbol": "star",
"stroke": "black",
"strokeWidth": 2,
"opacity": "opacity",
},
),
],
"grid": True,
"marginTop": 40,
"marginRight": 40,
"marginBottom": 40,
"marginLeft": 40,
"style": {"height": "400px"},
"title": Plot.js(
"`Dirichlet Mixture Model - Iteration ${$state.frame} of "
+ str(len(all_posterior_means) - 1)
+ "`"
),
"subtitle": "Cluster centers (★) with standard deviation (—) and data points (•)",
})
|
# Animation controls and legend with hover effects
Plot.html([
"div",
{"className": "p-4"},
[
"div",
{"className": "mb-4"},
Plot.Slider(
"frame",
init=0,
range=[0, len(all_posterior_means) - 1],
step=1,
label="Iteration",
width="100%",
fps=8,
),
],
[
"div",
{"className": "mt-4"},
Plot.js(
"""function() {
"""
+ frame_data_js
+ """
// Count assignments in current frame
const counts = {};
assignments.forEach(a => { counts[a] = (counts[a] || 0) + 1; });
"""
+ color_palette
+ """
// Sort clusters by weight, filter by minimum weight, and limit to top 10
const topClusters = Object.keys(weights)
.map(i => ({
id: parseInt(i),
weight: weights[i],
count: counts[parseInt(i)] || 0
}))
.filter(c => c.weight >= 0.01)
.sort((a, b) => b.weight - a.weight)
.slice(0, 10);
// Create placeholder rows for consistent height
const placeholders = Array(Math.max(0, 10 - topClusters.length))
.fill(0)
.map(() => ["tr", {"className": "h-8"}, ["td", {"colSpan": 3}, ""]]);
return [
"div", {},
["h3", {}, `Top Clusters by Weight (Iteration ${frame})`],
["div", {"style": {"height": "280px", "overflow": "auto"}},
["table", {"className": "w-full mt-2"},
["thead", ["tr",
["th", {"className": "text-left"}, "Cluster"],
["th", {"className": "text-left"}, "Weight"],
["th", {"className": "text-left"}, "Points"]
]],
["tbody",
...topClusters.map(cluster =>
["tr", {
"className": "h-8",
"style": {
"cursor": "pointer",
"backgroundColor": $state.hoveredCluster === cluster.id ? "#f0f0f0" : "transparent"
},
"onMouseEnter": () => { $state.hoveredCluster = cluster.id; },
"onMouseLeave": () => { $state.hoveredCluster = null; }
},
["td", {"className": "py-1"},
["div", {"className": "flex items-center"},
["div", {
"style": {
"backgroundColor": plotColors[cluster.id % 20],
"width": "24px",
"height": "24px",
"borderRadius": "4px",
"border": "1px solid rgba(0,0,0,0.2)",
"display": "inline-block",
"marginRight": "8px"
}
}],
`Cluster ${cluster.id}`
]
],
["td", {"className": "py-1"}, cluster.weight.toFixed(4)],
["td", {"className": "py-1"}, cluster.count]
]
),
...placeholders
]
]
]
];
}()"""
),
],
])
)
Out[6]:
For the interested reader, here's some exercises to try out to make this model better:
- Extend the model to infer the variance of the clusters by putting an inverse_gamma prior replacing the
OBS_VARIANCEhyperparameter and doing block-Gibbs on it using the normal-inverse-gamma conjugacy - Try a better initialization of the datapoint assignment: pick a point a use something like k-means and assign all the surrounding points to the same initial cluster. Iterate on all the points until they all have some initial cluster.
- Improve inference using SMC via data annealing: subssample 1/100 of the data and run inference on this, then run inference again on 1/10 of the data starting with the inferred choices for cluster means and weights from the previous trace, and finally repeat for the whole data.
Note that the model is still expected to get stuck in local minima (the clustering at the borders isn't great), and one way to improve upon it would be to use a split-merge move, via reversible-jump MCMC.