rocStats: Make ROC curve data.

In alkalait/gprege: Gaussian Process Ranking and Estimation of Gene Expression time-series

Description

Computes the points on an ROC curve by varying a threshold on the sorted outputs of the method in question.

Usage

  rocStats(outputs, groundTruthLabels, decreasing = TRUE)

Arguments

outputs	A numeric vector with the outputs of the evaluated method (e.g. likelihoods from gprege).
groundTruthLabels	A binary vector than contains the ground truth (e.g. which genes belong in the top-100 ground truth list).
decreasing	Logical. TRUE sorts outputs by decreazing order.

Value

stats

A list of numeric variables with the necessary statistics to compute an ROC curve, a precision-recall curve, etc.

Examples

  ## see compareROC.R
  data(FragmentDellaGattaData) ## Load demo data.
  rocStats(gpregeOutput$rankingScores, DGatta_labels_byTSNItop100, decreasing=TRUE)

alkalait/gprege documentation built on May 10, 2019, 9:18 a.m.