densur / pywoe Goto Github PK

Could you outline basic usage with some examples?

Direction of WoE transformed variables with respect to the target variable appears to be opposite of expected as suggested by coefficients of a Linear model. Please see attached example code with this post. Working with PIMA Indians diabetes data predicting Diabetes outcome (https://archive.ics.uci.edu/ml/datasets/pima+indians+diabetes) coefficient of "Age" variable is positive but WoE transformation is negatively related to Diabetes outcome. This is also true for other variables example Glucose.

Hi, is it possible to use this package to fill in the missing values instead of dropping them before using the package?

you probably will want to make this into a pip package and add requirements for the version of sklearn and other required packages

After running mywoe.bins where mywoe = woe.WoE() and mywoe.fit(x,y), my counts for goods and bads are reversed. My y variable denotes 1's as goods. Isn't this normal?