search

mljar / mljar-supervised

Python package for AutoML on Tabular Data with Feature Engineering, Hyper-Parameters Tuning, Explanations and Automatic Documentation
https://mljar.com
MIT License
3.02k stars 403 forks source link

Doubt - Does mljar autoML also take care of applying RandomScaler, minmaxScaler etc? #628

Closed puneeshkhanna closed 1 year ago

puneeshkhanna commented 1 year ago

Also will the model perform better then if we feed in the scaled data to the AutoML pipeline ?

pplonski commented 1 year ago

There shouldn't be any difference. AutoML can perform data scaling if needed.

puneeshkhanna commented 1 year ago

@pplonski - thanks for the quick replies. I m already trying mljar autoML on a kaggle running competition dataset with mode=compete and seeing exciting results. Is data scaling information revealed somewhere in the logs etc. Will try to figure out too.

pplonski commented 1 year ago

It should be available in framework.json file in the model directory.

I'm happy that you get good results :)