search

zmjones / edarf

exploratory data analysis using random forests
MIT License
68 stars 11 forks source link

drop binary factor #33

Closed zmjones closed 9 years ago

zmjones commented 9 years ago

when doing binary classification with type = "prob" in partial_dependence we should drop one factor level, (how we know which idk, an option?), change the prob attribute as FALSE and assign the column for the other level the name of the target variable. This will get handled like regression in plot_pd.

zmjones commented 9 years ago

fixed in c6d13d2cafdafa5f1815a6550352ed185a1a050f