Skip to contents

Score New Data With a Fitted Scorecard

Source: R/obwoe_scorecard.R
predict.obwoe_scorecard.Rd

Applies a fitted scorecard to new data. Reads only the binning, the coefficients and the scaling — never the raw fitted model — so a scorecard saved with saveRDS() keeps scoring after the engine's package is gone.

Usage

# S3 method for class 'obwoe_scorecard'
predict(
  object,
  new_data,
  type = c("score", "card", "link", "prob", "woe"),
  ...
)

Arguments

object

An "obwoe_scorecard" object.

new_data

A data.frame carrying the model's variables.

type

One of "score" (default), "card" (the sum of the integer points, i.e. what the deployed table gives), "link" (log-odds of the event), "prob" or "woe".

...

Ignored.

Value

A numeric vector, or a data.frame when type = "woe".

See also

obwoe_scorecard

Examples

# \donttest{
german <- read.csv(
  gzfile(system.file("extdata", "germancredit.csv.gz",
    package = "OptimalBinningWoE"
  )),
  stringsAsFactors = FALSE
)
german$default <- 1L - german$credit_risk
german$credit_risk <- NULL

sc <- obwoe_scorecard(german, target = "default", seed = 1)
#> Warning: 210 events for 14 variables (15.0 per variable): below the 20-events-per-variable rule of thumb, so the coefficients are unstable.
summary(predict(sc, german))
#>    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
#>   396.1   494.2   518.3   519.2   545.3   628.2 
# }