Direct using of categorical features

[sashulyak]

LightGBM can use categorical feature directly.

There is a categorical_feature parameter in LightGBM docs to deal with this behavior:
https://github.com/Microsoft/LightGBM/blob/master/docs/Parameters.md

It would be nice to add categorical feature support to pyLightGBM.

[gw00207]

I can see from the regression example here that you integerise categorical features and then include these with continuous variables:

le = LabelEncoder()

for col in df.columns.tolist():
    if 'cat' in col:
        df[col] = le.fit_transform(df[col])

I agree it would be nice to add categorical feature support to pyLightGBM rather than training on mixed continuous and integerised categories (which doesn't make sense to me)