myCurves: Model Performance Visualization
In xxzcool/scoremodel: Credit Scoring Modeling Toolbox

Description Usage Arguments Details See Also

myCurves is an integrated, flexible as well as easy to use function of visualization for model performance.

1
2
3

myCurves(ytrain, predict_ytrain, ytest, predict_ytest, ontest = TRUE,
  lift_bins = 10, P0 = 600, PDO = 50, color_scheme = 1,
  ifsave = TRUE)

`ytrain`	A vector corresponding the response variable encoded with 0 and 1 of train set.
`predict_ytrain`	A vector of predicted probability values of train set.
`ytest`	A vector corresponding the response variable encoded with 0 and 1 of test set.
`predict_ytest`	A vector of predicted probability values of test set.
`ontest`	Logical, decide plot on train or on test, default `TRUE`.
`lift_bins`	An integer, set the number of binnings, default `10`.
`P0`	An integer, the base score, default `600`.
`PDO`	An integer, the scale, default `50`.
`color_scheme`	An integer, choice the color scheme 1, 2 or 3, default scheme `1`.
`ifsave`	Logical, whether to save the chart as a file, default `TRUE`.

the result graph has four subgraphs which constitute a matrix, top left is ROC Curve, top right is Lift Figure, lower left is K-S Curve, and lower right is Distribution Figure of standard score.

Other model performance functions: Curve_Data, crossValidation, myks

xxzcool/scoremodel documentation built on May 4, 2019, 10:56 a.m.