摘要：AbstractHedonic price is the most popular determinant of housing market segmentation. Nonetheless, hedonic regression is often criticized for not being able to tackle heterogeneity of hedonic price of urban scapes, which creates structural instability in the regression. Spatial switching regression is used in this paper to develop localized hedonic regression ensuring constancy of the parameters within each localized regression. Central Tokyo and the outskirts are clearly divided into two segments. This research provides a new dimension to the existing methods of housing market segmentation. Additionally, this research also improves the predictive capability and model fitting statistics of hedonic regression.
