-
Notifications
You must be signed in to change notification settings - Fork 5
Open
Description
선형 모델의 한계를 넘어, 비선형적/복잡한 관계를 포착할 수 있는 비모수적 CATE 추정 방법론을 구현합니다.
(참고: https://matheusfacure.github.io/python-causality-handbook/landing-page.html)
- 선형 모델의 한계점 설명: 비모수적 접근의 필요성 강조 (Ref: Sec 18)
- 핵심 모델링 기법 추가
- 메타-러너 (Meta-Learners): S/T/X-Learner 구현 (Ref: Sec 21)
- 편향 제거 머신러닝 (Debiased ML): R-Learner 구현 (Ref: Sec 22)
- CATE 모델 평가 방법 추가: Uplift Curve 등 CATE 전용 평가지표 구현 (Ref: Sec 19)
Metadata
Metadata
Assignees
Labels
No labels