mrfse.exact

A penalized likelihood BIC-based to estimate Markovian neighborhoods.

Three algorithms for estimating a Markov random field structure.Two of them are an exact version and a simulated annealing version of a penalized maximum conditional likelihood method similar to the Bayesian Information Criterion. These algorithm are described in Frondana (2016) <doi:10.11606/T.45.2018.tde-02022018-151123>.The third one is a greedy algorithm, described in Bresler (2015) <doi:10.1145/2746539.2746631).

Rodrigo Carvalho

mrfse

Markov Random Field Structure Estimator

Rodrigo Carvalho 

Florencia Leonardi 

mrfse.exact function

<dl> <dt>a_size</dt>
<dd>Size of the alphabet.</dd> <dt>sample</dt>
<dd>A integer-valued matrix. Each value must belong range <code>0</code>
 and <code>a_size - 1</code>. Matrix has dimension <code>n</code> x <code>V</code>, where
 <code>n</code> is number of samples and <code>V</code> is number of nodes.</dd> <dt>c</dt>
<dd>The penalization constant. Must be positive.</dd> <dt>max_neigh</dt>
<dd>The maximum length of a candidate Markovian neighborhood. Must be
 non-negative and less than <code>ncol(sample)</code>.</dd>
</dl>

Arguments

Author

A Markov random field structure estimator — mrfse.exact

<dl>

 <dt>a_size</dt>
<dd>Size of the alphabet.</dd>

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

 <dt>c</dt>
<dd>The penalization constant. Must be positive.</dd>

 <dt>max_neigh</dt>
<dd>The maximum length of a candidate Markovian neighborhood. Must be
 non-negative and less than <code>ncol(sample)</code>.</dd>


</dl>

mrfse.exact: A Markov random field structure estimator

Description

Usage

Value

Arguments

Author

References

Examples