Select robust features by permutation importance¶

GitHubでこのAIアクセラレーターにアクセス

Machine learning models have biases using small data, and some industries (e.g., healthcare and manufacturing) lack labeled data. In light of this, a good approach is to select robust features to build models. This accelerator introduces an approach to select robust features, use multiple seeds for cross validation, add dummy features to compute the median permutation importance, and select the most robust dummy features.

このノートブックでは、以下のことを行う方法がまとめられています。

• DataRobotに接続する。
• Create multiple projects by multiple seeds and add dummy features.
• パフォーマンスが上位のモデルからアンサンブルを作成。
• Retrieve modeling permutation importance from the top-performing blender models.
• Remove features whose permutation importance is lower than dummy features.