The real estate industry plays an important role in China’s national economic system,especially in most small and medium-sized cities.For a long time,local finance has been heavily dependent on the development of the real estate industry.Local governments have formulated various policies to stimulate the investment and development of the real estate market,resulting in the rapid growth of housing prices and high deviation between housing prices and local social and economic development level.As the second-hand housing market can not directly bring large profits to the local finance,the local government only issued relevant decrees to strengthen administrative management,not fully aware of the importance of the second-hand housing market,ignoring the norms and cultivation of the second-hand housing market itself,resulting in a serious deviation from the development of the second-hand housing market and the new house market;Under the effect of market mechanism,the city of high quality residential and common low residential co-exist for a long time,urban residential market appear unbalanced development,inadequate phenomenon,social justice and social stability problems,and domestic and foreign scholars have focused mainly on the accessibility of characteristic variable difference between the influence of the unbalanced development of the housing market,for the scale effect of characteristic variables space research is insufficient,It is particularly necessary to strengthen the research cases of small and medium-sized cities with the same problems in order to enrich the theoretical samples of real estate market research.In this background,this article in fuqing city real estate center to register 2020 secondhand housing transaction data as an example,from the space measurement and statistical perspective,using spatial autocorrelation,spatial gradient interpolation,inverse distance space space such as measurement method,in-depth analysis of fuqing city spatial structure features of second-hand housing prices,confirmed that the housing price spatial distribution of non-stationary characteristics.Then draw on the analysis results of other scholars and the actual situation of Fuqing city,select the corresponding variables to build an analysis model.After comprehensive comparative analysis,this study adopts MGWR model to explore the influence mechanism of spatial action scale of variable factors,namely spatial heterogeneity.MGWR model not only fully considers the spatial attributes of data,but also calculates the spatial weight of variables through the kernel function to determine the bandwidth value of different variables,namely,the spatial scale of variable influence.Effectively avoid regression analysis by other interference factors.Finally,some suggestions are put forward according to the research results.The research results are as follows:(1)On the whole,the second-hand housing price in Fuqing is characterized by the coexistence of spatial autocorrelation and spatial heterogeneity,and the threshold of global spatial autocorrelation is 4.8km;"High-low" and "low-high" heterogenous spots appeared in local areas.The difference of spatial gradient symbol of housing price further shows that the spatial distribution of second-hand housing price in Fuqing has the characteristics of heterogeneity.(2)The research results show that the fitting effect of MGWR model regression analysis is better than GWR model and least square method(OLS)based characteristic price model,and the model is suitable for second-hand housing price research in small and medium-sized cities,which has important theoretical significance.(3)The analysis results of MGWR model show that all the regression coefficients of characteristic variables have positive and negative sign differences,and there are obvious spatial changes.The spatial action scales of characteristic variables are different.The analysis results are consistent with the actual situation of Fuqing city on the whole,and the research has important practical significance. |