본문 바로가기 주메뉴 바로가기
22대 대한민국 국회 국회정보길라잡이 국회의원검색
회원가입 로그인
HOME

정보검색

소장정보 검색

국회도서관 홈으로 정보검색 소장정보 검색

결과 내 검색

동의어 포함

고급검색

인명/단체명

저자정보 상세정보
인명/단체명을 입력하세요.

키워드

  • 대표어

  • 외국어

-

목차보기

Title Page

Abstract

Contents

Ⅰ. Introduction 10

Ⅱ. Related works 13

2.1. Causal Inference 13

2.2. Propensity Score 13

2.3. Covariate Balancing Propensity Score (CBPS) 13

2.4. Multinomial Logistic Regression 13

Ⅲ. Settings 15

3.1. Assumptions 15

3.2. Theorem 16

3.3. Problem Setting 16

Ⅳ. Method 18

4.1. Estimation of GPS 18

4.2. Balance Test 20

4.3. Modeling of treatment effect 22

Ⅴ. Application 23

5.1. Description of the data 23

5.2. Preprocessing 23

5.3. Balance test 25

5.4. Estimation of causal effect 26

Ⅵ. Conclusion 29

References 30

Appendix 34

6.1. Appendix A : Derivation of (3) 34

6.2. Appendix B: Derivation of (5) 34

List of Figures

Figure 1. Confidence band of potential outcome when treatment is Iodine. 26

Figure 2. Confidence band of potential outcome when treatment is Smoking. 26

Figure 3. Confidence band of potential outcome when treatment is Drinking. 27

Figure 4. Confidence band of potential outcome when treatment is Strength training. 27

Figure 5. Confidence band of potential outcome when treatment is Cardio training. 27

Figure 6. Confidence band of potential outcome when treatment is Sleeping time. 27

Figure 7. Confidence band of potential outcome when treatment is Medication compliance. 28

초록보기

 Thyroid function and related symptoms are known to be affected by daily habits such as consumption of iodine, alcohol, smoking or sleep deprivation. Using observational data, we estimate the causal effect of daily habits on hyperthyroidism symptoms. Daily habits are mostly continuous variables which call for generalized propensity score (GPS) methods. The GPS is the conditional distribution of the treatment given the covariates and is used to balance the distribution of confounders between the treated and untreated groups. Instead of maximizing the likelihood for estimating the parameters of the GPS, we minimize the covariance between treatment and confounders, following Fong et al.(2018). Furthermore in this work, we modify the estimating equation of Fong et al.(2018) so as to make balance of user level unobserved confounders as well as observed confounders. Finally, we provide a lifestyle coaching framework based on the causal inference analysis results.

추천서가 (다양한 추천 자료를 만나보세요)

권호기사보기

권호기사 목록 테이블로 기사명, 저자명, 페이지, 원문, 기사목차 순으로 되어있습니다.
기사명 저자명 페이지 원문 기사목차

권호

기사명 원문보기 기사목차
닫기