Second-hand Housing Batch Evaluation Model Of Zhengzhou Citv Based On Big Data And MGWR Model

The government report in March 2019 proposed "Improve the local tax system and steadily advance the real estate tax legislation".How to better evaluate the value of urban real estate is a major issue facing the reform of the real estate tax system.Establishing an accurate and efficient batch evaluation model is to evaluate houses.The core issue of value.Domestic batch evaluation started relatively late,and the research is mostly limited to the application and optimization of existing models.There are few researches on model innovation and basic data measurement methods.The research methods mainly include multiple regression analysis,characteristic price model,principal component analysis,and fuzzy cluster analysis.Existing research methods have certain limitations in application.Housing prices and characteristic factors have spatial non-stationarity in different locations of the city.Instead of presenting a simple linear relationship.Therefore,this study introduces a spatial measurement model to fully consider the spatial heterogeneity of housing prices and the spatial nonstationarity of characteristic factor variables,and combines it with big data information technology and geographic information systems to obtain basic data and model correction methods.A series of processes such as construction,spatial measurement calculation,result analysis and visualization,and batch evaluation model construction are sorted out and summarized,providing a complete set of theoretical foundations and ideas for building a batch evaluation model,and providing technical support for the reform of the real estate tax system.First,take the second-hand housing in the area within the Fourth Ring Road of Zhengzhou City as the research object,and use the train collector to crawl the secondhand housing transaction data of Zhengzhou City from the Anjuke and Fangtianxia second-hand housing website from 2010 to 2019.Using Python3.0 toolkit to crawl Zhengzhou city interest point data and road network,river and other vector data through the Gaode map API interface.Summarize and sort out the index system,and combine the economic development level,natural landscape,infrastructure layout and other conditions of Zhengzhou City to construct an index system that includes three types of characteristic factors:location,neighborhood,and architectural features;secondly,the least squares regression model is used(OLS model),spatial error model(SEM model),geographic weighted regression model(GWR model),geographic and time weighted regression model(GTWR model),multi-scale geographic weighted regression model(MGWR model)on the price and interpretation of second-hand housing in Zhengzhou Variables are studied and analyzed,and the barrier line and access point measurement method(BLAAP)is constructed to compare the model with the three measurement methods of buffer(BA),Euclidean distance(ED),and non-Euclidean distance(ND)previously studied.Research;Finally,based on the ArcGIS 10.2 platform,a batch evaluation model of second-hand housing in Zhengzhou is established based on spatial location to estimate housing prices.By entering the latitude and longitude of the housing price sample point,you can query the weight of each characteristic factor on the housing price,and calculate the value of the house.The research results show that:(1)The second-hand housing batch evaluation model in Zhengzhou constructed by the MGWR model corrected by the BLAAP measurement method has the highest fit,reaching 0.883,the Akaike Information Criteria(AIC)and Sigma estimate are the smallest,and 88.3%of the second-hand housing data samples The points can be well explained by the model.(2)With the rapid development of the big data Internet today,in the process of transforming the traditional city model to the smart city model,for mega cities with well-developed road networks and complete infrastructure,the basic data measurement method should be appropriately weakened.Consider road networks,rivers,Costs such as traffic lights should pay more attention to the difference in the impact of characteristic variables on housing prices due to different scale indexes and quality levels.(3)When the city undergoes major functional zone planning adjustments,the non-stationarity of the action scale has a greater impact on housing prices than the non-stationarity of time.The MGWR model is better than the GTWR model in explaining housing prices.(4)Zhengzhou city housing prices and the business service center have obvious synergistic effects;the kindergarten space has the strongest non-stationary effect,and the spatial distribution is extremely sensitive to urban housing prices.Government departments should optimize the spatial layout of the kindergartens;the residential environment of the community has gradually increased the impact of housing prices People are more inclined to choose livable areas with high greening rate and low population density.