R/predict.deepforest.R

In deepdive: Deep Learning for General Purpose

#' Predict Function for DeepForest
#'
#' @param object deepforest model object
#' @param newData pass dataframe for prediction
#' @param ... further arguments passed to or from other methods.
#' @return returns predictions vector or dataframe
#' @export
#' @export predict.deepforest
#' @importFrom  data.table rbindlist
#' @import rpart
#' @importFrom treeClust rpart.predict.leaves
#' @importFrom graphics barplot
#' @importFrom stats formula predict runif

predict.deepforest<-function (object, newData, ...)
{
  if (!inherits(object, "deepforest"))
    stop("Not a legitimate \"deepforest\" object")
  varCut = object$varcut
  modelGroup = object$chosenModels
  treeAugment = object$treeAugment



  predlist <- lapply(1:length(modelGroup), function(o) {
    if (treeAugment == F) {
      xFit <- data.frame(newData[, varCut[[o]]])
      names(xFit) <- varCut[[o]]
      deepnetPred <- predict.deepnet(modelGroup[[o]], xFit)

      if(object$modelType!='regress'){
        deepnetPred <- deepnetPred[, names(deepnetPred) != "ypred"]
      }

      for (i in names(deepnetPred)) {
        deepnetPred[, i] <- as.numeric(deepnetPred[,i])
      }
      deepnetPred
    }
    else {
      xFit <- data.frame(newData[, varCut[[o]]])
      names(xFit) <- varCut[[o]]
      augmentPred <- predict.deeptree(modelGroup[[o]],
                                      newData = xFit)
      if (object$modelType == "multiClass") {
        augmentPred <- augmentPred[, names(augmentPred) !=
                                     "ypred"]
        for (i in names(augmentPred)) {
          augmentPred[, i] <- as.numeric(augmentPred[,
                                                     i])
        }
      }
      augmentPred
    }
  })
  networkCount <- length(modelGroup)
  predarray <- array(unlist(predlist), c(nrow(newData), ncol(predlist[[1]]),
                                         networkCount))

  predDf <- data.frame(apply(predarray, 1:2, mean))
  names(predDf) <- colnames(predlist[[1]])
  if (object$modelType == "multiClass") {
    predDf$ypred <- stringr::str_remove_all(names(predDf),
                                            "y_")[max.col(predDf)]
    names(predDf) = stringr::str_remove_all(names(predDf),
                                            "y_")
  }
  return(predDf)
}