R/simulateBinaryResponseMat.R

In integIRTy: Integrating Multiple Modalities of High Throughput Assays Using Item Response Theory

simulateBinaryResponseMat <- function(a=a, b=b, theta=theta) {
  if(length(a)!=length(b)) stop('length of difficulty and discrimination should be equal!\n')
  nItem <- length(a)
  nExaminee <- length(theta)
	# P_ij for gene i in sample j
  P <- matrix(NA, nrow=length(theta), ncol=nItem)
  for(i in 1:nItem){
    P[, i] <- exp(a[i]*(theta-b[i]))/(1+exp(a[i]*(theta-b[i])))
  }
	# binary matrix
  X <- matrix(NA, nrow=nExaminee, ncol=nItem)
  for(i in 1:nItem){
    X[, i] <- rbinom(nExaminee, size=1, prob=P[, i])
  }
  if(!is.null(names(a))) colnames(X) <- names(a)
  if(!is.null(names(theta))) rownames(X) <- names(theta)
  X
}