Make with H2O.ai: Accuracy Masterclass Part 3 - Feature Selection Best Practices

Review feature selection best practices in our third accuracy masterclass. In this Make with H2O.ai session Data Scientist and Kaggle Grandmaster, Dmitry Larko walks through when to use feature selection and why, the main approaches to feature selection and the tradeoffs, and examples of feature selection techniques.

0:00 Introduction and Overview of Feature Selection
4:23 Filter Method - Variance Threshold
7:39 Filter Method - Univariate Analysis
10:15 Filter Method - Feature Pairwise Correlation
13:25 Embedded Method - L1 (Lasso) Regularization
17:53 Embedded Method - Random Forest Feature Importance
21:49 Wrapper Method - Recursive Feature Elimination (RFE)
25:53 Wrapper Method - Permutation Importance
31:13 Wrapper Method - Target Permutation Selection
35:15 Summary and Conclusion