free islet

https://towardsdatascience.com/introduction-to-shap-values-and-their-application-in-machine-learning-8003718e6827 Introduction to SHAP Values and their Application in Machine Learning Learn how the SHAP library works under the hood towardsdatascience.com

2023.09.04 - [AI,머신러닝] - XGBoost 피쳐 중요도 https://medium.com/the-artificial-impostor/feature-importance-measures-for-tree-models-part-i-47f187c1a2c3 Feature Importance Measures for Tree Models — Part I An Incomplete Review medium.com https://mljar.com/blog/feature-importance-in-random-forest/ Random Forest Feature Importance Computed in 3 Ways with Python The feature importance (variable importanc..

kaggle 튜토리얼 competition 격인 titanic 분석 https://www.kaggle.com/code/sunghwankang/titanic 최고 점수: 0.79904 (Random Forest) 피쳐: Pclass, Sex, Age, SibSp, Parch, Fare, Embarked Pclass, Sex, Embarked 피쳐 One-Hot 인코딩 numerical 컬럼 결측치 median 사용 XGBoost로 변경 후에 오히려 0.74로 점수가 낮아졌는데, overfitting을 완화하는 방향으로 hyperparameter를 조정하니 0.76으로 약간 상승. 최적 파라미터를 찾기 위한 grid search는 해보지 않음. EDA correlation hm_df = train_data[..