I have this given database:
I would like to predict column "y" using the columns "index_1","index_2","index_3" using the random forest classifier.
as you can see, column "size: does not have values for each observation.
My question is : Can I still use random forest classifier when I don't have data for all the observations, and if yes, is it ok?should I give value (e.g "noData") to the empty cells? will it harm the prediction? or maybe no need?