FpFnInv: FpFnInv

In tdhock/rankSVMcompare: Support Vector Machines for Ranking and Comparing

Description

Count false positive, false negative, and inversions. Each of these is considered an error. Both the real labels and the predictions should be in c(-1,0,1).

Usage

FpFnInv(true, pred)

Arguments

true	Actual labels.
pred	Model predictions.

Value

data.frame with 1 row and 5 columns: false.positive, false.negative, inversion, error, count.

Author(s)

Toby Dylan Hocking

Examples

values <- c(-1,0,1)
x <- expand.grid(true=values, pred=values)
(err <- FpFnInv(x$true, x$pred))
stopifnot(err$err == 6)
stopifnot(err$false.positive == 2)
stopifnot(err$false.negative == 2)
stopifnot(err$inversion == 2)
stopifnot(err$count == nrow(x))

tdhock/rankSVMcompare documentation built on May 31, 2019, 7:38 a.m.