Feature Engineering & Data Preprocessing

80% pekerjaan ML. Handling missing data, encoding (label/one-hot/target), scaling, feature creation, outlier handling. Studi kasus Kaggle dengan 800+ features.