public class Make_Woe
extends java.lang.Object
This class is about computing the Weights of Evidence for a categorical variable versus another binary one.
This computation is very important in the credit scoring field mainly for two reasons.
Generally the WoE gives a measure of the magnitude of difference between the interval bad rate and the average bad rate. (April 2012 Dr Mohammad Yousaf Shad ,Mian Khaliq Ur Rehman).
Computationally WoE is calculated as
WoE=ln(dsitr.Good%/dsitr.bad%)x100
Having computed the WoE for each different value of a categorical variable, then the Information value is:
IV=Ói(dsitr.Good%i-dsitr.bad%i) X ln(dsitr.Good%i/dsitr.bad%i)
As a rule of thumb when :
| Constructor and Description |
|---|
Make_Woe() |
| Modifier and Type | Method and Description |
|---|---|
double |
get_information_value() |
java.lang.Object[] |
get_woe_classes_ob() |
double[] |
get_woe_classes() |
java.lang.String[] |
get_woe_names_st() |
double[] |
get_woe_variable() |
void |
give_me_woe(java.lang.String[] target,
java.lang.String[] categorical)
This is the main method of the class that computes all the WoE related statistics
|
public void give_me_woe(java.lang.String[] target,
java.lang.String[] categorical)
This is the main method of the class that computes all the WoE related statistics
categorical - : An array of Strings where its distinct values will be used to compute the WoEtarget - : An array of Strings with exactly 2 different values (binary)public java.lang.String[] get_woe_names_st()
public double[] get_woe_classes()
public java.lang.Object[] get_woe_classes_ob()
public double get_information_value()
public double[] get_woe_variable()