spn.structure.leaves.parametric.Parametric.Categorical

class Categorical(p=None, scope=None)

Implements a univariate categorical distribution with $k$ parameters {pi_{k}}

representing the probability of the k-th category

The conjugate prior for these values would be a Dirichlet

p({pi_{k}}) = Dir(oldsymbollpha)

__init__(p=None, scope=None)

Methods

__init__([p, scope])

Attributes

k
name
parameters
type