GenerateDataForPerfCurve: Generate the values for performance curve (Precision and...

In mahfuz05062/FLEX_R: FunctionaL Evaluation of eXperimental perturbation (i.e. genome wide knockout) screens

Description

Generate the values for performance curve (Precision and Recall for example) to plot on x and y axis

Usage

GenerateDataForPerfCurve(
  value.predicted,
  value.true,
  neg.to.pos = FALSE,
  x.axis = "sensitivity",
  y.axis = "precision"
)

Arguments

value.predicted	Predicted value / Score (profile similarity / direct interaction)
value.true	Actual value (1/0 co-annotation for example)
neg.to.pos	assign TRUE to sort from neg to positive (pos -> neg is the default)
x-axis	what is the x-axis value you want? (TP, FP, TN, FN, precision, FPR, recall/sensitivity)
y-axis	what is the y-axis value you want? (TP, FP, TN, FN, precision, FPR, recall/sensitivity)

Value

PR -> (x: recall, y: precision), ROC -> (x: recall, y: fpr or 1 - specificity)

Examples

PerfCurve <- function(value.predicted, value.true, neg.to.pos = FALSE, type = 'PR', x.axis = 'sensitivity', y.axis = 'precision')

mahfuz05062/FLEX_R documentation built on May 27, 2021, 4:32 a.m.