scikit-learn 0.23 이 되면서 변한 점

최근에 scikit-learn이 0.23으로 버전이 업데이트 됐다! 업데이트된 점에 정리해준 글이 있어 공유한다. 

1. Python 3.6 or newer only 🐍

 

python 3.6 이상만 사용 가능하다!

 

2. Interactive pipeline graphics ⬇

 

pipeline을 시각화해주는 기능이 생겼다!! 

저 시각화해주는 것은 html로 되어있는데 참 좋았다. 

 

3. Poisson and gamma GLMs have arrived 🎉

linear_model.PoissonRegressor and linear_model.GammaRegressor 이 새로 나와서 이제 scipy를 쓸 필요가 없다

 

 

4. Calling fit() doesn’t show you everything 🚫

The fit() method returns the estimator you call it on. Now if you print the estimator, only the parameters that you changed from the defaults are

sklearn.set_config(print_changed_only=False)

5. n_features_in_ shows you how many features 🔢

n_features_in_

6. Easier sample dataset loading 🧭

pandas로 더 쉽게 만들어 주는 as_frame이 생겼다.

diabetes = load_diabetes(as_frame=True)
df_diabetes = diabetes.data

7. Avoid type hinting errors ⚠️

8. Improvements to experimental classes 🧪

HistGradientBoostingRegressor and HistGradientBoostingClassifier, the two LightGBM-inspired tree ensemble algorithms, are still experimental. They still need to be specially imported. However, they received a number of improvements. Same with IterativeImputer — it’s still experimental and has been improved.

 

9. Plays nicer with new pandas dtype 🐼

 

 

https://scikit-learn.org/stable/whats_new/v0.23.html

Version 0.23.0 — scikit-learn 0.23.0 documentation

 

 

https://towardsdatascience.com/9-things-you-should-know-about-scikit-learn-0-23-9426d8e1772c

9 Things You Should Know about Scikit-Learn 0.23 🎉