gen.evidence: Generate a sequence of evidence for a continuous node in a conditional gaussian bayesian network.

Description

The evidence for a node in an RHugin domain is generated as a linear sequence within the specified standard deviation from the marginal mean of the node. The evidence can be given as an input to absorb.gnbp

Usage

gen.evidence(gpfit, node, std = 2, length.out = 10, std.equal = TRUE)

Arguments

gpfit

an object of class "gpfit" obtained by using fit.gnbp

node

a character string specifying the name of a continuous node in the domain

std

a numeric value specifying the number of standard deviations of marginal distribution within which the evidence is generated. A numeric vector of length = number of nodes, must be specified when std.equal=FALSE.

length.out

a positive integer giving the desired length of the sequence.

std.equal

a logical value indicating whether same number of standard deviations should be used to generate evidence for all nodes. Default is TRUE.

Value

A matrix of evidence for each specified node

Details

Examples

Run this code

##Fit a network
data(mouse)
network<-fit.gnbp(mousegeno,mousepheno,alpha=0.1)

##Generate a sequence of evidence for a single node
evidence<-gen.evidence(network,node="Tlr12",std=2,length.out=20)