prophys.distributions

Probability distributions whose parameters are symbolic expressions.

Univariate families (Gaussian, Weibull, Gumbel, …), circular and directional distributions, multivariate and data driven distributions, and dependence structures (GaussianCopula, Correlation, Independent). Every family builds log_prob and, where available, cdf, quantile, reparameterized sample and moments as differentiable expressions, so their parameters can be fitted or depend on other model quantities.

Classes

Distribution	Base class of all probability distributions.
TransformedDistribution	Distribution of f(X) for X ~ base and an invertible transform f.
Gaussian	Normal distribution with mean mean and standard deviation sigma.
Weibull	Weibull distribution in scale and shape form: p(x) = (k/lam) (x/lam)^(k-1) exp(-(x/lam)^k).
Exponential	Exponential distribution with rate rate: p(x) = rate exp(-rate x) for x >= 0.
LogNormal	Log normal distribution: log X is Gaussian with mean mu and standard deviation sigma.
Gamma	Gamma distribution with shape concentration and rate rate.
Beta	Beta distribution on (0, 1) with shape parameters alpha and beta.
Uniform	Continuous uniform distribution on [low, high].
Gumbel	Gumbel (type I extreme value) distribution for maxima, with location loc and scale scale.
TruncatedGaussian	Gaussian with a log density renormalized to the interval [low, high].
Bernoulli	Bernoulli distribution over {0, 1} with success probability prob.
Categorical	Categorical distribution over k categories, indexed from 0, with probabilities probs.
Poisson	Poisson distribution over nonnegative integer counts with mean rate.
VonMises	Von Mises distribution on the circle (radians), the circular analog of a Gaussian.
WrappedGaussian	Gaussian wrapped onto the circle, for angular quantities in radians.
DirectionalMixture	Joint distribution of a direction and a conditional intensity, such as a wind rose.
MultivariateGaussian	Multivariate normal distribution parameterized by its mean and the Cholesky factor of its covariance.
Mixture	Finite mixture of component distributions with (possibly learnable) weights.
Empirical	Data driven distribution defined by a fixed sample set, such as observed wind speeds.
Histogram	Piecewise constant density over histogram bins.
GaussianCopula	Joint distribution coupling arbitrary marginals through a Gaussian copula.
Correlation	Bivariate Gaussian copula between two marginals, with correlation rho = tanh(rho_raw).
Independent	Joint distribution of independent marginals (a product density).

Functions

corr_cholesky_from_raw	Return the lower Cholesky factor of a correlation matrix built from an unconstrained vector.
correlated_uniforms	Map independent standard normals to correlated uniforms of a Gaussian copula.