R/covFactor.R

In TreeBUGS: Hierarchical Multinomial Processing Tree Modeling

# factor handling

# recodes factor labels into integer index
# returns a pure numeric data set covData and the old factor labels
covRecodeFactor <- function(covData, predType) {
  numCov <- ncol(covData)
  predFactorLevels <- vector("list", numCov)

  for (i in 1:numCov) {
    if (predType[i] != "c") {
      # store factor level labels:
      predFactorLevels[[i]] <- sort(unique(levels(as.factor(covData[, i]))))
      nLevel <- length(predFactorLevels[[i]])

      if (nLevel <= 1) {
        stop("Factor", colnames(covData)[i], "has only one factor level!")
      }

      # replace factor levels by an integer index:
      covData[, i] <- match(as.factor(covData[, i]), predFactorLevels[[i]])
      covData[, i] <- as.numeric(covData[, i])
    }
  }


  list(covData = covData, predFactorLevels = predFactorLevels)
}