برنامه نویسی با پایتون

Positive Unlabeled Learning is a binary classification task where the negative label is unknown. With that only the positive label is a 100% correct while the "negative" label can be either positive or negative. This statistical problem should be benchmarked against a normal supervised learning algorithm: CATBOOST. Choose three PU-learning approaches and run them against CATBOOST! در واقع می‌خواد مدل نظارتی رو با نیمه نظارتی مقایسه کنه و دیتاست هم از این سایت انتخاب بشه: Data set: Any data from the UCI machine learning reposity or used for catboost. ." It needs to be "PU-labeled