R/Simulate.R

In PathSelectMP: Backwards Variable Selection for Paths using M Plus

Simulate <-
function(n=1000,seedNum=1000,MissingYN=0,exampleNum=1){
if(exampleNum==1){
set.seed(seed=seedNum)
#example 1
W=rbinom(n,1,0.55)
Y=rnorm(n)
Z=rbinom(n,1,0.40)

V1=2.5*Z
V2=(1/(1+exp(-V1)))
V3=rbinom(n,1,V2)

X1=2.5*V3+W-0.5*Z
X2=(1/(1+exp(-X1)))
X3=rbinom(n,1,X2)

NewData=data.frame(X3,V3,W,Z,Y)
names(NewData)=c("X","V","W","Z","Y")
}

if(exampleNum==2){
#example 2
set.seed(seed=seedNum)
A=rbinom(n,1,0.55)
B=rbinom(n,3,0.50)
C=rnorm(n)
E=rbinom(n,5,0.25)
F=rnorm(n)
W=rnorm(n)

G=1.5*C+E+3.5
G1=(1/(1+exp(-G)))
G2=rbinom(n,3,G1)

H=2.5*B+2*C+1-2*rnorm(n,1,2)
H1=(1/(1+exp(-H)))
H2=rbinom(n,2,H1)

J=3*H-1+B+C
J1=(1/(1+exp(-J)))
J2=rbinom(n,2,J1)

NewData=data.frame(J2,H2,G2,A,B,C,E,F,W)
names(NewData)=c("J","H","G","A","B","C","E","F","W")
}

if(exampleNum==3){
#example 3
set.seed(seed=seedNum)
A=rbinom(n,1,0.55)+1
B=rbinom(n,3,0.50)+1
C=rnorm(n)+1
E=rbinom(n,5,0.25)+1
F=rnorm(n)+1

G=1.5*A+E-4*C+2*F
G1=(1/(1+exp(-G)))
G2=rbinom(n,3,G1)

H=2.5*B+2*C+1-2*rnorm(n,1,2)
H1=(1/(1+exp(-H)))
H2=rbinom(n,2,H1)

J=3*H-1
J1=(1/(1+exp(-J)))
J2=rbinom(n,2,J1)

NewData=data.frame(J2,H2,G2,A,B,C,E,F)
names(NewData)=c("J","H","G","A","B","C","E","F")
}


if(MissingYN==1){
for(i in 1:ncol(NewData)){
Ch=sample(1:n,n*.1)
NewData[Ch,i]=(-99)
}
}
return(NewData)}