R/calibration.R

In SAEval: Small Area Estimation Evaluation

calibration <- function (data,dir,sae,area)
  
{
  
  dir.name <- all.vars(dir)
  sae.name <- all.vars(sae)
  area.name <- all.vars(area)
  calib <- list()
  for (i in 1:length(area.name)) {
    data.s <- data[, c(area.name[i], dir.name, sae.name)]
    output5 <- data.frame(matrix(0, nrow = length(unique(data.s[,area.name[i]])), ncol = length(sae.name)))
    data.s[is.na(data.s[, dir.name]),dir.name ]<-0
    estimates5 <- aggregate(data.s, by = list(area1 = data.s[,area.name[i]]), sum)[, -c(1, 2)]
    output5 <- ((estimates5[, sae.name] - estimates5[, dir.name])/estimates5[,dir.name]) * 100
    output5 <- data.frame(output5)
    colnames(output5) <- sae.name
    calib[[i]] <- output5
  }
  names(calib) <- area.name
  all.output <- calib


}