randCPTs: A function to generate random CPTs for a given DAG

In kelvinyangli/mbmml: Learn Markov blankets using Minimum Message Length

Description

This function randomly generate conditional probability tables (CPTs) for a given DAG, with specified maximum arity and concentration parameter from a symmetric Dirichlet distribution. Values of variables are sampled from capital alphabet letters, A, B, ... The function has dependency on the library gtools on its rdirichlet() function.

Usage

randCPTs(dag, maxArity, beta, arities = NULL, debug = FALSE)

Arguments

dag	A DAG in the bnlearn format.
maxArity	The maximum arity for variables. That is, the maximum number of values that a variable can have.
beta	The concentration parameter is a symmetric Dirichlet distribution. When beta=1, it is the same as sampling from uniform distribution; when beta>1, CPT values are more likeliy to be equal since they are sampled from a center-peaked n-1 simplex; when beta<1, CPT values are extreme, since they are sampled from conners of the n-1 simplex.
arities	The vector of arities for all nodes. By default, it is NULL.
debug	A boolean argument to display detailed steps.

Value

The returned CPTs are stored in a list that matches bnlearn CPTs format, so that the function bnlearn::rbn() can be used to sample data.

kelvinyangli/mbmml documentation built on June 29, 2020, 3:12 a.m.