I have a label that has a natural ordering e.g. 0,1,2,3 where 0 is the worst activity measure and 3 is the best. For each label given by the model i need to also give the probability that it belongs to that label. I tried RandomforestRegression but when using predict_proba i got the error that it doesn't exist for random forest model.Am i right in thinking that in using sklearn packages for regression e.g. randomforest, decsion tree, xgboost i cannot give probability of it being in my label 0,1,2,3. E.G. If my regressor outputs 0.3 i may have a rule than if <0.3 = class 0. But i am unsure how i develop a probability here.

What models in python can i use to predict an ordinal target, i am new to this so any help appreciated. Should i just model as classification?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.