Package 'woe'

Weigth of Evidence

Description

Computes the Weight of Evidence and Information Value between Dependent and Independent variable.

Usage

woe(Data, Independent, Continuous, Dependent, C_Bin, Bad, Good)

Arguments

Data	: Name of Data Set
Independent	: Name of the Independent Variable
Continuous	: True if the variable is continuous, False if variable is Ordinal or Nominal
Dependent	: Name of the Targer Variable
C_Bin	: Count of Bins to be computed
Bad	: Which categorical variable do you want to be bad
Good	: Which categorical variable do you want to be good

Details

WOE

Value

Returns a DataSet with computed WoE and IV values on success or 0 on Failure

Note

"woe" shows the log-odds ratio between between Goods and Bads. In the Bivalued Dependenet variable, one value represents Goods and others are bads. In Detail with an Example: Let Dependent varaible be ATTRITED (0,1) and Independent variable be TENURE where, 1-Attrited, 0-Non Attrited. If I wish to check WOE and IV of Tenure with ATTRITED to know if Tenure has an effect in getting attrited, Then good would be 1 and bad=0. If I wish to check WOE and IV of Tenure with ATTRITED to know if Tenure has an effect in not getting attrited, Then good would be 0 and bad=1.

Author(s)

Sudarson Mothilal Thoppay

Examples

woe(Data=mtcars,"cyl",FALSE,"am",10,Bad=0,Good=1)
woe(Data=mtcars,"mpg",TRUE,"am",10,Bad=0,Good=1)

---