collElimination2: Collinearity Elimination by Giving Variables' IV

In xxzcool/scoremodel: Credit Scoring Modeling Toolbox

Description

collElimination2 is a recursive function, which will recursivly eliminate the collinearity X variables with smaller IV. Different from collElimination, it must provide parameter IVdf, so need not recalculate every variable's IV for speed up.

Usage

collElimination2(df, IVdf, cor_critical = 0.8)

Arguments

df	A dataframe only with Xs and Y variable, all the X variables must have been binned, such as after woe-encoded.
IVdf	A dataframe of every X variable's IV.
cor_critical	Specifies the correlation coefficient threshold for extracting, default 0.8.

Details

for collinearity elimination, recursion is necessary to avoid mistakenly-deleted.

Value

A dataframe after elimination

See Also

Other collinearity elimination functions: collElimination

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