Samples#

class cuqi.samples.Samples(samples, geometry=None, is_par=True, is_vec=True)#

An object used to store samples from distributions.

Parameters:
  • samples (ndarray) – Contains the raw samples as a numpy array indexed by the last axis of the array.

  • geometry (cuqi.geometry.Geometry, default None) – Contains the geometry related of the samples

__init__(samples, geometry=None, is_par=True, is_vec=True)#

Methods

__init__(samples[, geometry, is_par, is_vec])

burnthin(Nb[, Nt])

Remove burn-in and thin samples.

ci_width([percent])

Compute width of the pointwise credibility intervals of the samples

compute_ci([percent])

Compute pointwise credibility intervals of the samples.

compute_ess(**kwargs)

Compute effective sample size (ESS) of samples.

compute_rhat(chains, **kwargs)

Compute rhat value of samples given list of cuqi.samples.Samples objects (chains) to compare with.

diagnostics()

Conducts diagnostics on the chain (Geweke test).

hist_chain(variable_indices, *args, **kwargs)

Plots samples histogram of variables with indices specified in variable_indices.

mean()

Compute mean of the samples.

median()

Compute pointwise median of the samples

plot([sample_indices])

Plots one or more samples.

plot_autocorrelation([variable_indices, ...])

Plot the autocorrelation function of one or more variables in a single chain.

plot_chain([variable_indices])

plot_ci([percent, exact, plot_envelope_kwargs])

Plots the credibility interval for the samples according to the geometry.

plot_ci_width([percent])

Plot width of the pointwise credibility intervals of the samples

plot_mean(*args, **kwargs)

Plot pointwise mean of the samples

plot_median(*args, **kwargs)

Plot pointwise median of the samples

plot_pair([variable_indices, kind, marginals])

Plot marginals using a scatter, kde and/or hexbin matrix.

plot_std(*args, **kwargs)

Plot pointwise standard deviation of the samples

plot_trace([variable_indices, exact, ...])

Creates a traceplot of the samples consisting of 1) a histogram/density plot of the samples and 2) an MCMC chain plot.

plot_variance(*args, **kwargs)

Plot pointwise variance of the samples

plot_violin([variable_indices])

Create a violin plot of the samples.

std()

Compute pointwise standard deviation of the samples

to_arviz_inferencedata([variable_indices])

Return arviz InferenceData object of samples for the given variable indices

variance()

Compute pointwise variance of the samples

Attributes

Ns

Return number of samples

funvals

Returns a new Samples object of sample function values.

geometry

is_vec

parameters

If self.is_par is False, returns a new Samples object of sample parameters by converting the function values to parameters.

shape

Returns the shape of samples.

vector

Returns a new Samples object of samples in vector form.