plearn.struct: Infer network structure for mixed types of random variables.

In equSA: Learning High-Dimensional Graphical Models

Description

Learning graphical model structure for mixed types of random varaibles based on p-learning algorithm. Each variable in the dataset can be either binary or Gaussian distributed.

Usage

plearn.struct(data, gaussian.index = NULL, binary.index = NULL,
alpha1 = 0.1, alpha2 = 0.02, restrict = FALSE, score.only=FALSE)

Arguments

data	The data matrix, of dimensions nxp. Each row is an observation vector and each column is a variable.
gaussian.index	The index vector of Gaussian nodes. The default value is NULL. If not specified, the system will automatically determine the index for each variable.
binary.index	The index vector of binary nodes. The default value is NULL. If not specified, the system will automatically determine the index for each variable.
alpha1	The significant level of parent and children screening in p-learning algorithm. The default value is 0.1.
alpha2	The significant level of moral graph screening in p-learning algorithm. The dafault value is 0.02.
restrict	Should edge restriction applied? (logical). If TRUE, we assume that there should be no edge among binary variables. The default is FALSE.
score.only	If TRUE, it only reports z-scores for all pair of variables. The default is FALSE.

Details

This is the function that implements the p-learning algorithm for learning the undirect network structure for mixed types of random variables.

Value

A list of two objects.

Adj	The estimated adjacency matrix of undirect network.
score	The estimated z-scores for all pair of variables.

Author(s)

Bochao Jia and Faming Liang

References

Jia, B., and Liang, F. (2018) Joint Estimation of Restricted Mixed Graphical Models. manuscript.

Examples

 library(equSA)
 data.graph <- DAGsim(n = 200, p = 100, type="AR(2)", p.binary = 50)$data
 plearn.struct(data.graph, alpha1 = 0.1, alpha2 = 0.02)

equSA documentation built on May 6, 2019, 1:06 a.m.