gainsChartPerfCurve: Produce a cumulative performance graph based on gains...

In dnegrey/miscTools: R package with miscellaneous tools

Description

gainsChartPerfCurve produces a cumulative performance graph based on a mt_gainsChart object. When two such objects are supplied, the function produces a combined graph, where the first object (xb) is labeled as "Build" and the second object (xv) is labeled as "Validate".

Usage

gainsChartPerfCurve(xb, xv = NULL, cb = "#009DDC", cv = "#77BF30",
  cr = "#666666", xLabel = "Cumulative Total Percent",
  yLabel = "Cumulative Outcome Percent")

Arguments

xb	mt_gainsChart object
xv	mt_gainsChart object
cb	character string; line color for xb (valid color)
cv	character string; line color for xv (valid color)
cr	character string; random/reference line color (valid color)
xLabel	character string; x-axis label
yLabel	character string; y-axis label

Value

A plot_ly visualization of a mt_gainsChart object, showing the cumulative percent of total Y (cumPctY for "binary"/"continuous"; cumPctYC for "combined") by quantile ( yhatBin) percent.

See Also

gainsChart, plot_ly

Examples

# pull in sample scored data frame
x <- modelSampleScored
# filter to validation subset
x <- split(x, x$ValidateFlag)
# gains chart binary
y <- lapply(x, function(z){gainsChart(z$TargetFlag, z$pTargetFlag)})
# cumulative performance graph
gainsChartPerfCurve(y[[1]], y[[2]])
# gains chart continuous
y <- lapply(x, function(z){
    gainsChart(z[z$TargetFlag, "TargetValue"], z[z$TargetFlag, "pTargetValue"])
})
# cumulative performance graph
gainsChartPerfCurve(y[[1]], y[[2]])
# gains chart combined
y <- lapply(x, function(z){
    gainsChart(z$TargetFlag, z$pTargetFlag*z$pTargetValue, z$TargetValue)
})
# cumulative performance graph
gainsChartPerfCurve(y[[1]], y[[2]])

dnegrey/miscTools documentation built on May 3, 2019, 2:57 p.m.