dataFormat: Put data into a format usable by the main functions

In andreamrau/GBNcausal: Causal estimation for Gaussian Bayesian networks

Description

Put the data into a format usable by the main functions of the package.

Usage

dataFormat(x, int.nodes = c(), int.means = c())

Arguments

x	matrix - Observations or simulated observations from a GBN.
int.nodes	matrix - 0 indicates that there is no intervention on the node. 1 indicates that the node is knock-out. Only those two values are supported by the MCMC.GBN function.
int.means	matrix - Values of the knock-out : if int.nodes = 1 then x = int.means. Only 0 is currently supported by the MCMC.GBN function.

Value

data is a list of x, int.nodes and int.means with the same colnames/rownames. If there is no colnames/rownames in x, names given by the function are of the form "N#" where # is a number.

See Also

dataCreate

Examples

p<-10

W1=1*upper.tri(matrix(0,p,p))
m1=rep(0,10)
sigma1=rep(0.1,10)

X=simulGBN(2*p,m1,sigma1,W1,seed = 1990)
X=rbind(X,simulGBN(p,m1,sigma1,W1,int=1,int_data=0,seed = 1990)) 
X=rbind(X,simulGBN(p,m1,sigma1,W1,int=9,int_data=0,seed = 1990)) 

intnode=matrix(0,4*p,p)
intnode[(2*p+1):(3*p),p-9]=1
intnode[(3*p+1):(4*p),p-1]=1

colnames(X)=colnames(intnode)=paste("N",1:p,sep="")

dataFormat(X,intnode) 

andreamrau/GBNcausal documentation built on May 12, 2019, 3:34 a.m.