서울시 토지가격 영향요인의 비선형성과 공간적 이질성 탐색- GeoShapley의 적용 -

초록

Land prices are the core element shaping property prices, land use patterns, and public finance. Furthermore, access to diverse urban services and facilities shapes local variations in land prices. Traditional hedonic price models have extensively investigated the determinants of land prices by assuming linear relationships, failing to identify submarket segmentation. To overcome these limitations, this study combines the automated machine learning and GeoShapley methods to better model the nonlinear and spatially varying effects of different factors on residential and nonresidential land prices. Our analysis confirms that spatial location is the most influential factor shaping residential and nonresidential land prices. Our empirical models also find nonlinear effects and spatial heterogeneity in effects of access to schools, hospitals, shopping facilities, parks, public transit, and urban centers on land prices. These results inform real estate policy by considering the importance of spatial heterogeneity and threshold effects, thus extending the classical theories of urban property markets by adding price-setting mechanisms in specific submarkets.

키워드