Aspiring Data Science
327 subscribers
390 photos
10 videos
6 files
1.46K links
Заметки экономиста о программировании, прогнозировании и принятии решений, научном методе познания.
Контакт: @fingoldo

I call myself a data scientist because I know just enough math, economics & programming to be dangerous.
Download Telegram
#featureselection #kuhn

Читаю по рекомендации товарища книжку по ML. В главе по FS есть задание, мимо которого не смог пройти ) Надо будет потестить на нём Диогена. А возьмётся кто-то из читателей потестить на этом примере алгоритмы sklearn/mlxtend?
#uplift #kuhn

Понравилась идея matched samples в аплифт-моделировании.

"Another approach could be to use more sophisticated sampling techniques to create an appropriate training set. For the table above, it is impossible to contact and to not contact the same customer. However, in medical research, this problem is often faced when evaluating a new treatment against an existing therapy. Here, clinical trials sometimes use matched samples. Two subjects are found that are nearly identical and are randomized into treatment groups. The idea is that the only differentiating factor is the treatment, and the patient response can be estimated more accurately than without matching. The important idea here is that the subjects are no longer the experimental unit. The matched pair itself becomes the primary data point in the analysis."
#novelty #outlier #kuhn

Вот такая простая, но перспективная идея по самодельному детектору новизны.