Research On Second-hand Housing Valuation Model Of Lanzhou

With the rapid development of economy,economic activities related to real estate are becoming more and more frequent.Especially in urban areas,where there is less land available for development,So the second-hand housing gradually entered the considera-tion of the people.and the demand for real estate information is increasing.Therefore,if second-hand housing prices can quickly and accurately evaluated,there is positive effec-tive for developers to develop real estate and residentsThe paper considers the ordinary housing in the second-hand housing of five main urban areas in Lanzhou City of Gansu Province as the research object.Firstly,The three important factors,regional,construction and neighborhood are investigated,and 21 char-acteristic variables are selected to establish the index system;Then the data are collected by python 2 to crawl websites related data of second-hand housing in Anju Ke and Fang Tianxia,and WebGIS program implements the query of the neighborhood peripheral fa-cilities,which is more accurate than the questionnaire survey;Thereafter single evaluation models are established,such as random forest,SVR,boosting and so on,to forecast and evaluate the housing price separately.The results show that the error of the random forest prediction is the lowest,but the prediction effect is not necessarily the best at all obser-vation sample points.Therefore,an ensemble Stacking learning algorithm,is explored as the combination model to improve the forecasting performance.The results show that the standardized mean square error(NMSE)0.147,reduced the relative random forest regression 27.07%.So the ensemble Stacking learning algorithm has the best prediction effect,with lower error and good generalization performance.Finally,to partition the unit area price of second-hand housing into three levels,the high-grade second-hand housing,the middle ones and the low-sheltered ones.Upon the criterion of variable importance of random forest and OWA operators,three different types of second-hand houses were studied separately according to their important variables.The research can provide some constructive suggestions for house buyers and builders.