if I plot the binning table, I get something like the following
so basically, any numeric value should get a woe of 1.27 and any null should get woe of -0.3.
however, if I retrieve the unique values of the binned_X0, I get [0, 1.27] so the woe of the numeric values are, as expected 1.27 but the woe of the null is given a zero which should have been assigned to special values of which I have none.
is this how it should behave or this is really unexpected?
guillermo-navas-palencia
commented
4 months ago
Hi, I am using BinningProcess to get the woe of a feature like X0 below.
binning = BinningProcess(X0.columns.to_list(), n_jobs=-1) binned_X0 = binning.fit_transform(X0, y_train, metric='woe')
if I plot the binning table, I get something like the following
so basically, any numeric value should get a woe of 1.27 and any null should get woe of -0.3.
however, if I retrieve the unique values of the binned_X0, I get [0, 1.27] so the woe of the numeric values are, as expected 1.27 but the woe of the null is given a zero which should have been assigned to special values of which I have none.
is this how it should behave or this is really unexpected?
Best regards, Afshin