cnFindBIC: Find Network by BIC
In sdnet: Soft-Discretization-Based Bayesian Network Inference

Description Usage Arguments Details Value Author(s) See Also Examples

This is a model selection routine that finds a network in a set of networks using the BIC criteria.

1	cnFindBIC(object, numsamples)

`object`	A `list` of `catNetworkNode` objects or `catNetworkEvaluate` or `dagEvaluate`
`numsamples`	The number of samples used for estimating `object`

The function returns the network with maximal BIC value from a list of networks as obtained from one of the search-functions cnSearchOrder, cnSearchSA and cnSearchSAcluster. The formula used for the BIC is log(Likelihood) - 0.5*Complexity*log(numNodes).

A catNetwork object with optimal BIC value.

N. Balov

cnFindAIC, cnFind

  cnet <- cnRandomCatnet(numnodes=12, maxpars=3, numcats=2)
  psamples <- cnSamples(object=cnet, numsamples=10)
  nodeOrder <- sample(1:12)
  nets <- cnSearchOrder(data=psamples, pert=NULL, 
	maxParentSet=2, maxComplexity=36, nodeOrder)
  bicnet <- cnFindBIC(object=nets, numsamples=dim(psamples)[2])
  bicnet