DistributionCollection.sample (function)

def sample(self, *args, **kwargs)

Sample from multiple distributions with random seeds automatically determined, but applied to distributions of the same underlying multivariate distribution automatically.

For each unique BaseDistribution.uniqueid in the distributions in dists a random seed will be generated and applied to BaseDistribution.sample for all distributionis in dists which share that same uniqueid value. By doing so, any BaseMultivariateDistribution which samples from the same underlying multivariate distribution (but for a different BaseMultivariateDistribution.dimension), will be correctly sampled to account for the covariance/correlation between parameters, but all other 1-D BaseDistribution objects will be sampled with their own independent random seeds.

Arguments

• *args: all positional arguments are sent to BaseDistribution.sample for each item in dists.
• cache_sample (bool, optional, default=True): whether to cache the sampled values for subsequent calls to DistributionCollection.pdf, DistributionCollection.logpdf, etc.
• **kwargs: all keyword arguments are sent to BaseDistribution.sample for each item in dists. Note: seed is forbidden and will raise a ValueError.

Returns

• (list): list of samples, in same order as DistributionCollection.distributions.

Raises

• ValueError: if seed is passed.