The Construction And Application Of Price Index Of Second-hand House Under The Background Of Big Data

In China,the real estate market is changing and reorganizing.Due to the inadequacy of new houses in location and living facilities,people's purchase strategies have shifted to the second-hand housing market,which also means that the real estate market in China will be transferred from the new housing transaction market to the second-hand housing transaction market.Now people's demand for the information of real estate becomes more and more urgent in this situation.On the one hand,the accurate,timely,and complete information of second-hand housing price index provides decision-making references for government,real estate developers,and the both sides of the transaction.On the other hand,it plays a vital role in the development of the real estate market.Therefore,this article combines the characteristics of big data and second-hand housing market,using Random Forest algorithm to build a set of second-hand housing price index calculation models with high prediction accuracy,high processing speed,and good stability.When calculating the price index of second-hand housing,the methods used by domestic and foreign researchers are Cost Input Method,Weighted Average Method,Hedonic Price Method,Repeat Sale Method and Pooled Method.In calculation process,these traditional methods are usually affected by human subjective factors,and the number of samples is limited.With insufficient comparable cases,big errors could be caused,so it cannot reflect the actual changes of the second-hand housing market.In recent years,domestic and foreign researchers have applied statistical modeling method to analyze the real estate market,and achieved good results.This thesis uses Random Forest algorithm to build the second-hand house price index model.Random Forest is a new machine learning method based on the statistical learning theory.It solves the fitting of the nonlinearity and limited data,and have better generalization and the steady frame.Thus,Random Forest is suitable for the building of second-hand housing price index model.In this thesis,we select four communities in Jiangbei District of Chongqing for empirical research.Using the Random Forest algorithm in R software to determine optimal values of parameters(mtry=5,ntree=700),to calculate second-hand housing prices index,compare forecast results,and finally establish a second-hand house price index model.We can find that the value of the second-hand housing price predicted by the Random Forest model is very close to the real second-hand housing price indexin the research area from the study findings.The Random Forest model has a good prediction result.And the model can serve the government macroeconomic regulation and control,foreign investment and the consumer purchasing more accurately,timely and objectively,so the model has a certain practical value.