강한상관이 존재하는 보조정보를 이용한 회귀 추정

다국어 초록 (Multilingual Abstract)

Regression estimation can be improved survey precision by incorporating auxiliary information, but highly correlated auxiliary variables often cause multicollinearity and unstable coefficient estimation. This study studies regression estimation using highly correlated auxiliary variables under complex survey designs and proposes weighted principal component regression (WPCR) and weighted ridge regression (WRR) as alternatives to the generalized regression estimator (GREG). Monte Carlo simulations across varying correlation structures evaluate bias, standard deviation, RMSE, and absolute relative bias (ARB). The results show that WPCR and especially WRR provide more stable and accurate estimation than GREG in highly correlated settings.

목차 (Table of Contents)

• 제1장 서 론 ·········································································································· 1
• 제2장 이론적 배경 ······························································································ 3
• 제1절 회귀모형과 회귀계수 추정의 기본 개념 ····································· 3
• 제2절 다중공선성의 개념과 회귀추정에 미치는 영향 ························· 5
• 제3장 회귀추정(Regression Estimation) ··················································· 7
• 제1절 가중 최소제곱 회귀(WLS) ······························································ 7
• 제2절 가중 주성분 회귀(WPCR) ····························································· 10
• 제3절 가중 능형 회귀(WRR) ··································································· 13
• 제4장 데이터 분석 ··························································································· 16
• 제1절 자료 소개 및 시뮬레이션 방법 ···················································· 16
• 제2절 모의실험 ······························································································ 19
• 제3절 실제 데이터 분석 ············································································· 25
• 제5장 결 론 ········································································································ 27
• 참 고 문 헌 ·································································································· 29
• ABSTRACT ······································································································ 31