For a logistic regression credit risk model, Weight of Evidence (WOE) encoding is applied to a high-cardinality categorical variable. Why is WOE particularly suited for logistic regression?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Here's the deal — b is correct because WOE = log(Distribution of Events / Distribution of Non-Events) for each category, which maps to the log-odds scale of the logistic regression model. This linearity in log-odds makes the encoded predictor directly compatible with the logistic regression framework.
Full explanation below image
Full Explanation
B is correct because WOE = log(Distribution of Events / Distribution of Non-Events) for each category, which maps to the log-odds scale of the logistic regression model. This linearity in log-odds makes the encoded predictor directly compatible with the logistic regression framework. A is right that WOE reduces to one numeric column but that is not the primary statistical advantage. C is partially true but not WOE's defining benefit. D is wrong; WOE is not required by PROC LOGISTIC.