mrfse.ci {mrfse}

Bresler's non-binary Markov random field structure estimator

Description

A greedy algorithm to estimate Markovian neighborhoods.

Usage

mrfse.ci(a_size, sample, tau, max_degree=ncol(sample)-1)

Arguments

a_size

Size of the alphabet.

sample

A integer-valued matrix. Each value must belong range 0 and a_size - 1. Matrix has dimension n x V, where n is number of samples and V is number of nodes.

tau

A hyperparameter. See references.

max_degree

The maximum length of a candidate Markovian neighborhood. Must be non-negative and less than ncol(sample).

Value

A list filled with estimated Markov neighborhood for each graph vertex

Author(s)

Rodrigo Carvalho

References

Guy Bresler. 2015. Efficiently Learning Ising Models on Arbitrary Graphs. In Proceedings of the forty-seventh annual ACM symposium on Theory of Computing (STOC '15). Association for Computing Machinery, New York, NY, USA, 771–782. DOI:https://doi.org/10.1145/2746539.2746631

Examples

library(mrfse)
a_size = c(0, 1)
s = matrix(sample(a_size, size=1000, replace=TRUE), ncol=5)
mrfse.ci(length(a_size), s, 0.2)