TruncatedNormal

class pmrf.distributions.TruncatedNormal(loc: Array, scale: Array, low: Array, high: Array)

Bases: AbstractProbDistribution

Truncated Normal distribution with loc, scale, low, and high parameters.

Initializes a Truncated Normal distribution.

Arguments:

• loc: Mean of the untruncated distribution.

• scale: Standard deviation of the untruncated distribution.

• low: Lower bound of the truncation.

• high: Upper bound of the truncation.

cdf(value: Array) → Array

See Distribution.cdf.

entropy() → Array

Calculates the Shannon entropy (in nats).

icdf(value: Array) → Array

See Distribution.icdf.

kl_divergence(other_dist, **kwargs) → Array

Calculates the KL divergence to another distribution.

Raises:

NotImplementedError: The KL divergence between arbitrary truncated normal distributions lacks a stable general analytical solution.

log_cdf(value: Array) → Array

See Distribution.log_cdf.

log_prob(value: Array) → Array

See Distribution.log_prob.

log_survival_function(value: Array) → Array

See Distribution.log_survival_function.

mean() → Array

Calculates the mean.

median() → Array

Calculates the median.

mode() → Array

Calculates the mode.

sample(key: Key[jaxlib._jax.Array, '']) → Array

See Distribution.sample.

sample_and_log_prob(key: Key[jaxlib._jax.Array, '']) → tuple[Array, Array]

See Distribution.sample_and_log_prob.

stddev() → Array

Calculates the standard deviation.

survival_function(value: Array) → Array

See Distribution.survival_function.

variance() → Array

Calculates the variance.

property event_shape: tuple[int, ...]

Shape of event of distribution samples.

high: Array

loc: Array

low: Array

scale: Array

property support: tuple[Array, Array]

See Distribution.support.