InverseWishart#
- class hybrid_aif.distributions.InverseWishart[source]#
Bases:
ModuleInverse-Wishart posterior over a full noise covariance.
Psiis(*parent_shape, d, d)(positive definite),nuis(*parent_shape,)with the propriety floorν > d − 1(the mean exists fromν > d + 1).- stats(scatter, count)[source]#
Package one tick’s statistic: full residual scatter
(..., d, d)and responsibility mass(...,).
- sample(key)[source]#
One noise draw as an
R_cholarray, via the Bartlett decomposition ofW ~ Wishart(ν, Ψ⁻¹)andR = W⁻¹.
- __init__(Psi, nu, Psi0=None, nu0=None)#