Hi there, I am working with a Logistic regression model and for categorical
variables with many categories I am using Weight-Of-Evidence (WOE) instead
of individual dummies for each category. Does anyone know if when I do this
I would have to transform all the other variables into their correspondant
WOE or the regression could be run with only a subset of variables in this
form?
Many thanks in advance!
Regards
Leo.
|