Skip to content

Repository files navigation

『실무로 통하는 인과추론 with 파이썬』

이 저장소는 한빛출판사에서 출간한 한국어판 "실무로 통하는 인과추론 with 파이썬"에서 참조하는 각종 자료 및 소스 코드와 예제 데이터를 담고 있습니다. 이 책은 다음 서점에서 절찬리에 판매중입니다(교보문고, 예스24, 알라딘, 한빛미디어)

img

책 소개

독자 여러분은 "상관관계는 인과관계가 아니다"라는 말을 많이 들어보셨죠? 이 책은 상관관계가 왜 인과관계와 다른지 그리고 데이터 과학자의 시각에서 인과추론의 기초부터 심화에 이르기까지의 내용을 다양한 실무 사례를 바탕으로 설명합니다. 나아가 인과추론 방법론에 그치지 않고 실무에서의 데이터 분석가 또는 데이터 과학자가 고민하는 수준의 실험 결과의 신뢰성 및 추론 부분까지 다루고 있습니다. 여러분이 인과추론을 처음 접하셨다면, 내용을 바탕으로 파이썬 실습 위주로 진행해볼 수 있습니다. 그리고 여러분이 데이터 분석가/과학자라면 사례를 중심으로 여러분의 도메인에 접목시켜 이론과 사례를 함께 학습할 수 있습니다.

주요 내용은 다음과 같습니다.

  • 인과추론의 기본 개념과 활용법 익히기
  • 인과추론과 편향의 관계 이해하기
  • 인과추론으로 비즈니스 문제 해결하기
  • 인과추론으로 고객을 시간에 따라 관찰하기
  • 인과효과가 실험 대상마다 다를 수 있는 이유 학습하기

역자 소개

네오플을 거쳐 크래프톤의 데이터 분석가로 일하고 있습니다. 게임 업계에서 쌓은 커리어를 기반으로 <던전앤파이터>, <뉴스테이트 모바일>, <배틀그라운드 모바일> 등 다양한 장르의 게임에서 데이터 분석과 실험을 통해 유저 경험을 개선하는 데 기여했습니다. 비영리 데이터 사이언스 커뮤니티인 가짜연구소에서 인과추론팀을 운영 중입니다. 마테우스 파쿠레의 웹북 「Causal Inference for The Brave and True」를 한국어로 번역하는 작업을 주도했습니다 (블로그 / Github)

가짜연구소 인과추론팀은 2022년부터 데이터를 통한 문제해결력을 높이고자 인과추론을 함께 학습하고 있습니다. 한국어 자료가 많지 않은 인과추론을 많은 분이 쉽게 접하실 수 있도록 기여하고자 하는 마음으로, 가짜연구소에서 인과추론 이야기와 실험 및 조직문화에 대한 이야기를 이어나가고 있습니다. 이 책의 번역 작업에는 인과추론팀 김소희, 김성수, 김상돈, 김준영, 남궁민상, 박시온, 최은희, 정호재, 홍성철이 함께 참여했습니다.

감수자 소개

미국 조지아대학교 테리 경영대학(Terry College of Business)에서 경영정보시스템 조교수로 일하고 있다. 디지털기술의 사회적, 환경적 영향에 대한 실증연구를 해오며 인과추론 방법론을 통해 사회현상과 기업활동에서의 원인과 결과를 분석하는 일들을 수행하고 있다. 인과추론의 저변을 확대하기 위해 매년 여름 를 조직해왔고, 유튜브 채널 <인과추론의 데이터과학>을 운영중이다.

챕터별 자료

번역서 온보딩을 위한 강의 & 블로그

참고 도서

책 논문 및 자료 모음

1장-인과추론 소개

챕터내용자료 링크
1.6 인과효과 식별하기더 알아보기Causal Inference and Data Fusion in Econometrics

2장-무작위 실험 및 기초 통계 리뷰

챕터내용자료 링크
2.2 A/B 테스트 사례시뮬레이션 데이터와 실제 데이터 비교A Randomized Assessment of Online Learning
2.4 가장 위험한 수식하워드 웨이너의 유명한 글The Most Dangerous Equation
2.5 추정값의 표준오차MIT의 통계학 입문 강좌Introduction to statistics
2.6 신뢰구간실제 사례: 코로나19 백신의 효과Efficacy and Safety of the mRNA-1273 SARS-CoV-2 Vaccine
2.8 p 값실제 사례: 실제 사례: 대면 강의 vs. 온라인 강의Is It Live or Is It Internet? Experimental Estimates of the Effects of Online Instruction on Student Learning
2.10 표본 크기 계산더 알아보기A/B Testing Intuition Busters: Common Misunderstandings in Online Controlled Experiments

3장-그래프 인과모델

챕터내용자료 링크
3.7.2 랜덤화 재해석민감도 분석과 부분 식별Making Sense of Sensitivity: Extending Omitted Variable Bias
3.8.1 충돌부를 조건부 설정더 알아보기DAGitty, A Crash Course in Good and Bad Controls

4장-유용한 선형회귀

챕터내용자료 링크
4.1 선형회귀의 필요성OLS 연구Difference-in-differences with variation in treatment timing, Interpreting Ols Estimands When Treatment Effects are Heterogeneous: Smaller Groups, Contamination Bias in Linear Regressions
4.7.5 평균 제거와 고정효과패널데이터를 사용한 인과추론 논문On the Pooling of Time Series and Cross Section Data
4.9 중립 통제변수잡음 제거 기법Improving the Sensitivity of Online Controlled Experiments by Utilizing Pre-Experiment Data

5장-성향점수

챕터내용자료 링크
5.1 관리자 교육의 효과민감도 분석과 부분 식별Estimating Treatment Effects with Causal Forests: An Application
5.3.3 성향점수 매칭민감도 분석과 부분 식별Why Propensity Scores Should Not Be Used for Matching
5.3.5 역확률 가중치의 분산실제 사례: 인과적 콘텍스트 밴딧A Contextual-Bandit Approach to Personalized News Article Recommendation
5.5 이중 강건 추정이중 강건 추정량 주석An Introduction to the Augmented Inverse Propensity Weighted Estimator
5.5.2 결과 모델링이 쉬운 경우더 알아보기Comment: Performance of Double-Robust Estimators When “Inverse Probability” Weights Are Highly Variable
5.6 연속형 처치에서의 일반화 성향점수연속형 처치Causal inference with a continuous treatment and outcome: Alternative estimators for parametric dose-response functions with applications

6장-이질적 처치효과

챕터내용자료 링크
6.7 누적 이득 곡선더 알아보기Nubank fklearn, Causal inference and uplift modeling a review of the literature, Empirical Analysis of Model Selection for Heterogeneous Causal Effect Estimation
6.8 목표 변환더 알아보기Intelligent Credit Limit Management in Consumer Loans Based on Causal Inference
6.9.2 이진 결과더 알아보기Causal Classification: Treatment Effect Estimation vs. Outcome Prediction

7장-메타러너

챕터내용자료 링크
7.1 이산형 처치 메타러너인과추론 라이브러리EconML 공식문서, CausalML 공식문서
7.1.1 T 러너더 알아보기Metalearners for estimating heterogeneous treatment effects using machine learning
7.2.1 S 러너S 러너의 편향Double/Debiased/Neyman Machine Learning of Treatment Effects
7.2.2 이중/편향 제거 머신러닝트리 기반 및 신경망 러너Nonparametric Estimation of Heterogeneous Treatment Effects:From Theory to Learning Algorithms, Learning Representations for Counterfactual Inference

8장-이중차분법

챕터내용자료 링크
8.2 표준 이중차분법실제 사례: 최저임금과 고용Minimum Wages and Employment: A Case Study of the Fast-Food Industry in New Jersey and Pennsylvania
8.2.5 추론더 알아보기When Should You Adjust Standard Errors for Clustering?, An Exact and Robust Conformal Inference Method for Counterfactual and Synthetic Controls
8.3 식별 가정더 알아보기When Is Parallel Trends Sensitive to Functional Form?
8.3.2 비기대 가정과 SUTVA공간적 파급 효과Difference-in-Differences Estimation with Spatial Spillovers
8.3.5 피드백 없음순차적 무시 가능성Causal Inference with Time-Series Cross-Sectional Data: A Reflection
8.3.6 이월 효과와 시차종속변수 없음순차적 무시 가능성When Should We Use Unit Fixed Effects Regression Models for Causal Inference with Longitudinal Data?
8.6.3 최종 결과이중 강건 이중차분법Doubly robust difference-in-differences estimators
8.7 처치의 시차 도입이중 강건 이중차분법Difference-in-differences with variation in treatment timing
8.7 처치의 시차 도입실제 사례: 개발도상국의 고등교육과 성장Higher Education Expansion, Labor Market, and Firm Productivity in Vietnam
8.7.1 시간에 따른 이질적 효과더 알아보기Estimating dynamic treatment effects in event studies with heterogeneous treatment effects, Two-Way Fixed Effects, the Two-Way Mundlak Regression, and Difference-in-Differences Estimators, Difference-in-Differences with multiple time periods

9장-통제집단합성법

챕터내용자료 링크
9.4 표준 통제집단합성법통제집단합성법에 대한 가정Using Synthetic Controls: Feasibility, Data Requirements, and Methodological Aspects
9.7 추론더 알아보기An Exact and Robust Conformal Inference Method for Counterfactual and Synthetic Controls
9.8.4 통제집단합성법과 이중차분법원본 합성 이중차분법Synthetic Difference-in-Differences
9.9 요약일반화 통제집단합성법Generalized Synthetic Control Method: Causal Inference with Interactive Fixed Effects Models, A Bayesian Alternative to Synthetic Control for Comparative Case Studies
9.9 요약실제 사례: causalimpact 라이브러리CausalImpact Github

10장-스위치백 실험

챕터내용자료 링크
10.2.2 무작위 탐색최적화Synthetic Controls for Experimental Design, Designing Experiments with Synthetic Controls
10.2.2 무작위 탐색다른 실험 목표Synthetic Design: An Optimization Approach to Experimental Design with Synthetic Controls
10.3.5 강건한 분산더 적은 가정으로 m 찾기Design and Analysis of Switchback Experiments

11장-도구변수

챕터내용자료 링크
11.9.4 밀도 불연속 테스트최적화Manipulation of the Running Variable in the Regression Discontinuity Design: A Density Test
11.10 요약최적화Does Compulsory School Attendance Affect Schooling and Earnings?

12장-더 배울 내용

챕터내용자료 링크
12.1 인과관계 발견인과관계 발견Causal Discovery Toolbox: Uncover causal relationships in Python
12.3 인과적 강화학습콘텍스트 밴딧Contextual Bandits in a Survey Experiment on Charitable Giving: Within-Experiment Outcomes versus Policy Learning
12.3 인과적 강화학습미국 경제학회 웹캐스트2022 Continuing Education Webcasts
12.4 인과 예측미국 경제학회 웹캐스트2019 AEA Continuing Education Webcasts
12.5 도메인 적응Concept DriftLearning under Concept Drift: A Review

참고할 링크

About

책 "실무로 통하는 인과추론 with 파이썬" (한빛미디어, 2024)의 내용을 정리

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Used by

Contributors

Languages