Heterogeneous Logit Model 을 이용한 광역수도권의 주택유형 및 교통수단 선택에 관한 연구

• 영문명: A Simultaneous Choice of Housing and Transportation Mode in the Metropolitan Capital Area by a Heterogeneous Logit Model
• 발행기관: 한국지역개발학회
• 저자명: 이진원(JinWon Lee) 김원수(WonSoo Kim) 주현태(HyunTai Choo)
• 간행물 정보: 『한국지역개발학회지』제13권 제1호, 163~183쪽, 전체 21쪽
• 주제분류: 사회과학 > 지역개발
• 파일형태: PDF
• 발행일자: 2001.04.30

영문 초록

This study aims to improve the statistical precision of aggregate probabilistic choice model. Models are applied in the choice of housing and transportation mode affected by the rapid change of the structure in the metropolitan capital area. The data are 3,176 effective ones sampled from population of the capital area. The major findings are as follows. First, a choice of housing and transportation mode is discussed employing Multinominal Logit Model as a basic model. The results of estimation are quite well in accordance with those of existing studies. Second, Nested Logit Model encompassing Multinominal Logit Model is analyzed to test the validity of Multinominal Logit Model. Nested Logit Model is an appropriate model since an Inclusive Value is significantly different from 1. The signs of most variables are desirable and the coefficients are statistically significant. In applying Nested Logit Model, Limited Information Maximum Likelihood method (LIML) has usually been used. LIML is, however, less efficient than Full Information Maximum Likelihood method (FIML). This study obtains better efficiency by using FIML. Third, as the highlight of this study, two strong assumptions of Nested Logit Model are relaxed; the constant similarity and homoskedastic error terms. The validity of Heterogeneous Logit Model with the flexible similarity and heteroskedastic error terms is confirmed. This study shows that Heterogeneous Logit Model could enhance the statistical precision of probabilistic choice model.