R/conv.ann.R

In ConvertPar: Estimating IRT Parameters via Machine Learning Algorithms

conv.ann<-function(small.data, train.data, model="2PL",layers=1,learningrate=NULL,treshold=0.01){

  p<- colMeans(train.data)

  total.score<-rowSums(train.data)

  r<-c()
  for(i in 1:ncol(train.data)){

    r[i]<-cor(train.data[,i],total.score)
  }

  mod<-mirt(train.data,model=1, temtype = model)

  par<-coef(mod,IRTpars=TRUE,simplify=TRUE)

  if(model=="2PL"){
    a <- par[[1]][,1]
  }

  else if(model=="Rasch"){
    a<-rep(1,length(r))
  }

  b <- par[[1]][,2]

  datapars<-cbind(a,b,r,p)

  if(model=="2PL"){

    modela<-neuralnet(formula = a ~ p+r ,data = datapars,hidden = layers)
    modelb<-neuralnet(formula = b ~ p+r ,data = datapars,hidden=layers)
  }

  else if(model=="Rasch"){

    modelb<-neuralnet(formula = b ~ p+r ,data = datapars,hidden = layers)
  }

  else{
    stop("Please select one of following options as a string : Rasch, 2PL (Default model is 2PL) ")
  }

  p<-colMeans(small.data)

  total.score<-rowSums(small.data)
  r<-c()

  for(i in 1:ncol(small.data)){
    r[i]<-cor(small.data[,i],total.score)
  }

  params<-cbind(p,r)

  if(model=="2PL"){
    a_pre<- predict(modela,params)
    b_pre<- predict(modelb,params)

    predicted_irt_params<-cbind(a_pre,b_pre)
    colnames(predicted_irt_params)<-c("a.pre","b.pre")
  }

  if(model=="Rasch"){

    b_pre<- predict(modelb,params)
    predicted_irt_params<-as.data.frame(b_pre)
    colnames(predicted_irt_params)<-c("b.pre")
  }

  train.data.parameters<-datapars
  output<-list(predicted_irt_params,train.data.parameters)
  names(output)<-c("Predicted IRT Parameters","Item Parameters of Training Data")

  return(output)
}