Causal ML パッケージã¨å¦ã¶ Meta-Learner - Qiitaã¯ã˜ã‚㫠仕事関係㧠Uplift Modeling ã«ã¤ã„ã¦èª¿ã¹ã¦ã„ãŸã‚‰ã€CATE (Conditional Average Treatment Effect) ã«ãŸã©ã‚Šç€ãã¾ã—ãŸã€‚ CATE 㯠ATE (Average Treatment Effect) ã‚’ã‚る特徴é‡ã§æ¡ä»¶ä»˜ã‘ãŸã‚‚ã®ã§ã€ATE ãŒ"å¹³å‡çš„ãª"処置効果を算出ã—ã¦ã„ã‚‹ã®ã«å¯¾ã—ã€åŠ¹æžœã¯å„属性 (特徴é‡) ã«ã‚ˆã£ã¦å¤‰ã‚ã‚‹ã¯ãšã§ã‚ã‚‹ã¨ã„ã†è€ƒãˆã®ã‚‚ã¨ã€éžå‡è³ªæ€§ (heterogeneity) を織り込んã å½¢ã§ã®å‡¦ç½®åŠ¹æžœã‚’算出ã—ã¦ã„ã¾ã™ã€‚ $$ATE:=E[Y(1)-Y(0)]$$ $$CATE:=E[Y(1)-Y(0)|X=x]$$ ã“ã“ã§ã€$Y(1)$ã€$Y(0)$ ã¯æ½œåœ¨çš„çµæžœå¤‰æ•°ã€$X=x$ ã¯ã‚る特徴é‡ã¨ãªã‚Šã¾ã™ã€‚ CATEã€ã™ãªã‚ã¡å€‹äººã‚„セグメントレベルã§ã®å‡¦ç½®åŠ¹æžœã‚’推定ã™ã‚‹ã“ã¨ãŒã§ãã‚Œã°ã€å‡¦ç½®åŠ¹æžœãŒãƒ—ラスã®äººã«ã®ã¿ã‚
About CausalML — causalml documentation
機械å¦ç¿’を用ã„ãŸå› 果推論ã®ãŸã‚ã®æœ€æ–° Python パッケージ "CausalML" ã®ç´¹ä»‹ - YuRAN-HIKO
Uberå¾¹åº•ç ”ç©¶ -å› æžœæŽ¨è«–ã«ã‚ˆã‚‹ãƒžãƒ¼ã‚±ãƒ†ã‚£ãƒ³ã‚°æœ€é©åŒ–ç·¨-｜Tack
uber/causalml: Uplift modeling and causal inference with machine learning algorithms
EconML/notebooks/Metalearners Examples.ipynb at main · py-why/EconML
コードã¨ã‚µãƒ³ãƒ—ルデータを公開 · iwanami-datascience/vol3@51bdebf
gitcdn.link
GitHub - itamarcaspi/EconML-with-R: How to use EconML within R%26date%3D20110511)
医療ã‹ã‚‰å®‡å®™ã¸ トップ分æžä¼šç¤¾ | iAnalysis(アイアナ)
{{#tags}}- {{label}}
{{/tags}}