R/PerformanceList.R

In simonmoulds/lulcc2: Land Use Change Modelling in R

#' @include class-PerformanceList.R
NULL

#' Create a PerformanceList object
#'
#' This function uses different measures to evaluate multiple
#' \code{ROCR::\link[ROCR]{prediction}} objects stored in a
#' \code{\link{PredictionList}} object.
#'
#' @param pred an object of class PredictionList
#' @param measure performance measure to use for the evaluation. See
#'   \code{ROCR::\link[ROCR]{performance}}
#' @param x.measure a second performance measure. See
#'   \code{ROCR::\link[ROCR]{performance}}
#' @param \dots additional arguments to \code{ROCR::\link[ROCR]{performance}}
#' 
#' @seealso \code{\link{performance}}, \code{\link{PredictionList}}
#' @return A \code{PerformanceList} object.
#'
#' @export
#'
#' @references Sing, T., Sander, O., Beerenwinkel, N., Lengauer, T. (2005).
#' ROCR: visualizing classifier performance in R. Bioinformatics
#' 21(20):3940-3941.
#'
#' @examples
#'
#' ## see lulcc2-package examples

PerformanceList <- function(pred, measure, x.measure="cutoff", ...) {
    perf <- performance(pred@prediction, measure, x.measure, ...)
    auc <- performance(pred@prediction, measure="auc")
    auc <- sapply(auc, function(x) unlist(slot(x, "y.values")))
    auc <- as.numeric(formatC(auc, digits=4, format="f"))
    out <- new("PerformanceList",
                performance = perf,
                auc = auc,
                categories = pred@categories,
                labels = pred@labels)
}

setGeneric("performance", function(prediction.obj, ...)
           standardGeneric("performance"))

setMethod("performance", signature(prediction.obj = "list"),
          function(prediction.obj, measure, x.measure="cutoff", ...) {
              perf <- list()
              for (i in 1:length(prediction.obj)) {
                  perf[[i]] <- ROCR::performance(prediction.obj[[i]], measure, x.measure, ...)
              }
              perf
          }
)

## # Create ROCR performance objects
## #
## # A wrapper function for \code{ROCR::\link[ROCR]{performance}} (Sing et al,
## # 2005) to create \code{performance} objects from a list of \code{prediction}
## # objects.
## #
## # param prediction.obj a list of \code{ROCR::\link[ROCR]{prediction}} objects 
## # param measure performance measure to use for the evaluation. See
## #   \code{ROCR::\link[ROCR]{performance}}
## # param x.measure a second performance measure. See
## #   \code{ROCR::\link[ROCR]{performance}}
## # param \dots additional arguments to \code{ROCR::\link[ROCR]{performance}}
## #
## # seealso \code{ROCR::\link[ROCR]{prediction}},
## # \code{ROCR::\link[ROCR]{performance}}
## # return A list of \code{performance} objects.
## #
## # export
## # rdname performance.rocr
## # 
## # references Sing, T., Sander, O., Beerenwinkel, N., Lengauer, T. (2005).
## # ROCR: visualizing classifier performance in R. Bioinformatics
## # 21(20):3940-3941.

## setGeneric("performance", function(prediction.obj, ...)
##            standardGeneric("performance"))

## # rdname performance.rocr
## # aliases performance,list-method
## setMethod("performance", signature(prediction.obj = "list"),
##           function(prediction.obj, measure, x.measure="cutoff", ...) {
##               perf <- list()
##               for (i in 1:length(prediction.obj)) {
##                   perf[[i]] <- ROCR::performance(prediction.obj[[i]], measure, x.measure, ...)
##               }
##               perf
##           }
## )