Using Geographically Weighted Regression to Redefine Housing Submarkets

• Main Author: 陳力綸
• Other Authors: 張金鶚
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/05492689651402633176

碩士 === 國立政治大學 === 地政研究所 === 101 === Housing market is a bundle of houses with various characteristics, and it can be disaggregated into submarkets by different definitions which has been an important study issue in a decade. Housing submarkets defined by a priori manners such as geographical boundaries or certain characteristics were proved to be a non-optimal way. Recent studies tried to define submarkets with statistics methods such as principal component analysis and cluster analysis. There is some agreement that using these statistics methods are more reasonable to divide submarkets than those a priori ways, but the lack of concern for spatial dependence and spatial heterogeneity is a major problem to define optimal submarkets. Besides, there have been studies arguing about the continuity and discontinuity in housing submarkets. And the optimal number of housing submarkets is still a rare yet important topic. Hence, this study is focused on these issues. This study redefines submarkets using geographically weighted regression for the Taipei City. Use the regression results to compose a homogeneity index and divide houses into different submarkets. Moreover, compare the newly-defined submarkets with a priori submarkets and find an optimal number for housing submarkets.