Research The Characteristic Price Of Second-hand Housing Based On Space Hedonice Model

With the continuous acceleration of China's urbanization process,the real estate industry has gradually become one of the important industries of China's national economy.However,with the rising house prices,ordinary buyers are under pressure.Therefore,the topic of housing prices has received a lot of attention from the government,society and academia.Therefore,it is of great practical significance to deeply study the price change law of real estate industry commodities.In the early residential price research,the traditional feature price model was used to explore.With the development of spatial econometrics,the academic community has gradually paid attention to the special nature of commercial housing.Therefore,the study of residential prices has gradually begun to explore using spatial measurement models.This paper will also use the spatial econometric model to explore the residential feature price,and compare it with the traditional feature price model to select the optimal model.The paper data was obtained and pre-processed by reptile technology,and finally 990 second-hand housing transaction data were screened.Then through the literature research combined with the actual situation of second-hand housing in Nanchang,19 indicators were selected and the corresponding data were quantified.The empirical part is mainly composed of three parts: first,calculate the characteristic price model under different functional forms,and compare and select the functional form with the highest fitting degree;secondly,the distribution of residential prices in Nanchang through the trend analysis tool in ArcGIS Analyze and further obtain Moran'I index,Geary's C index and Moran scatter plot to verify whether there is spatial correlation and residential price space agglomeration form.Third,use spatial lag model and spatial error model to analyze sample data.And compared with the traditional feature price model to get the optimal model.The following conclusions are drawn: 1.Of the three functional forms of the feature price model,the full logarithmic form has the highest R2 of 0.815,greater than 0.792 for the semi-logarithmic form R2 and 0.791 for the linear form R2.2.The price of second-hand housing in Nanchang was obtained by ArcGIS to obtain Moran'I and Geary's C,which were 0.126 and 0.378 respectively,indicating that the residential price was not randomly distributed and there was obvious agglomeration.According to the Moran scatter plot study,Nanchang residential prices mainly show “high-high agglomeration” and “low-low agglomeration”.3.The spatial lag model and the spatial error model are compared with the traditional eigenvalue model.The spatial lag model and the spatial error model R2 are 0.832 and 0.863 respectively higher than the traditional feature price model of 0.815,so the spatial econometric model has a better explanation.Sex.At the same time,the log-likelihood estimate of the spatial lag model is 371.487 smaller than the 491.423 of the spatial error model,so the spatial error model is the optimal model.4.The significant variables of residential price in the spatial error model are building type,level,bedroom living room,area,age,decoration degree,subway,CBD,property management fee,leisure and entertainment,and key schools.Among them,the residential price is inversely proportional to the characteristic variable age and CBD,and the others are proportional.