AveragePrecision.R

# compute average precision, given a matrix of true labels and a matrix of predicted probabilities.
# true_labels: Matrix of true labels, columns corresponding to labels and rows to instances.
# predictions: Matrix of probabilities predicted by a classifier.

AveragePrecision<-function(true_labels, predictions){
  return(mldr::average_precision(true_labels, predictions))
  # if(nrow(true_labels) != nrow(predictions) | ncol(true_labels) != ncol(predictions)){
  #   stop("labels and predictions should be in the same format")
  # }
  # result = 0
  # M=ncol(true_labels)
  # N=nrow(true_labels)
  # 
  # result<-sapply(seq(nrow(true_labels)), function(x){
  #   tmp<-rbind(true_labels[x,], predictions[x,])
  #   tmp<-tmp[,order(tmp[2,],decreasing = T)]
  #   res<-tmp[1,][1]
  #   
  #   return(res)
  # })
  # return(sum(result)/N)
}