Closed ilyes495 closed 7 years ago
The "target dataframe" (actually it must be a pandas Serie) must contain the labels of training dataset only.
Also if you're dealing with a classification problem, you need to have the target encoder python object in the directory "save" (LabelEncoder
fitted) and saved as "target_encoder.obj"
I will add more details to the docs tomorrow ! Thanks for reporting this issue !
I have a question regarding the format of target dataframe which is passed to fit_predict(), does it contain the labels of training dataset or for test dataset as well ? and how it should be formatted? I think it's not clear if someone has to process the data himself before passing it to the predictor.
Thanks