perf_df: perf_df

In wkostelecki/ezplot: Functions for Common Chart Types

perf_df

Description

shows classification performance statistics as a table

Usage

perf_df(fitted, actual, quantiles = NULL)

Arguments

fitted	A character value. Evaluates to a numeric column.
actual	A character value. Evaluates to a logical or binary column.
quantiles	Number of quantiles to show. If NULL, uses distinct values of fitted for the cutoffs rather than showing quantiles.

Value

A data.frame summarizing binary classification performance:

• quantile: fitted value quantile (only if !is.null(quantile)

• cutoff: fitted value cutoff

• fp: false positives

• tp: true positives

• tn: true negatives

• fn: false negatives

• pp: positive predictions

• np: negative predictions

• ipp: group positive predictions

• ifp: group false positives

• itp: group true positives

• rpp: rate of positive predictions

• acc: accuracy

• fpr: false positive rate

• tpr: true positive rate

• fnr: false negative rate

• tnr: true negative rate

• prec: precision

• clift: lift

• ilift: group lift

• f1: f1 measure

• ks: Kolmogorov-Smirnov statistic

• auc: area under ROC curve

• aucpr: area under PR curve

Examples

perf_df(mtcars$mpg, mtcars$am)
perf_df(mtcars$mpg, mtcars$am, quantiles = 4)
perf_df(mtcars$mpg, mtcars$am, quantiles = 8)
perf_df(mtcars$mpg, mtcars$am, quantiles = 10)
perf_df(mtcars$wt, mtcars$am==0)

wkostelecki/ezplot documentation built on Feb. 1, 2024, 10:58 p.m.