Skip to content

markov.md

Latest commit

 

History

History
26 lines (23 loc) · 934 Bytes

markov.md

File metadata and controls

26 lines (23 loc) · 934 Bytes
CurrentModule = Oscar
CollapsedDocStrings = true
DocTestSetup = Oscar.doctestsetup()

Discrete random variables

The joint probability distribution of random variables X_1, \ldots, X_n is given by a tensor of order n. If the random variable X_i takes d_i states, the tensor is of format d_1 \times \cdots \times d_n and consists of non-negative real numbers p_{x_1 \cdots x_n}, for all choices x_i \in [d_i], which sum to 1. The functions below deal with the ambient polynomial ring in these p variables, special forms in them like marginals, and conditional independence ideals.

markov_ring
ring(R::MarkovRing)
random_variables(R::MarkovRing)
gens(R::MarkovRing)
state_space(R::MarkovRing, K=random_variables(R))
parameter_ring(M::DiscreteGraphicalModel{Graph{Undirected}, T}; cached=false) where T
marginal(R::MarkovRing, K, x)
indexed_ring(R::Ring, varnames; kw...)