China's New Commercial Residential Housing Price Index Research Based On Autoregressive Mixed Effect Model

House price index plays an extremely important role in the real estate market research and analysis of the macroeconomic guidance.International house prices experiences the first generation method represented by the median and simple weighted average method and the second generation method represented by the Laplace weighting method.The third generation method i.e.homogeneity is gradually becoming the mainstream of the international house prices index method.The domestic new commercial residential housing price indexes mostly use the simple average method,such as the median method and the weighted average method.They are the equal of the first and second generation method of abroad.The behindhand theory leads to shortcomings in accuracy.However,the concept,the characteristics,the idea of residents and related data for housing are significant different between domestic and foreign country.The third generation index establishment methods do not completely suitable for domestic situation.So it is meaningful to establish the related theory and method of new commercial residential housing price index based on Chinese features with the basic theory of the third generation model.First of all,based on the proposed virtual trading data generation method based on floor,this article puts forward virtual trading data generation method based on direction.We also present the data fusion processing method for the virtual data.Xiang yang mortgage data in 2012 is used to generate virtual repeated trading data.Then we use BMN model,Case-Shiller model,OFHEO model to calculate the prices and compare them with statistics index.Come to the conclusion that the virtual repeated sales model is applicable to calculate China's new housing price index.Second,based on the mixed effect model and consider the impact of standard height of the floor and total height and the interaction between them,we establish the interactive mixed effect model.Use the Xiang yang 2012 mortgage data for empirical analysis and compare the result of two models.Come to the conclusion that interactive mixed effects model is a better choice.Finally,on the basis of mixed effect model,autoregressive mixed effect model is put forward.The model makes full use of the virtual repeated trading data and single data and considers the relation of previous deal with the current transaction.Use Xiang Yang 2012 mortgage data for calculation.By comparing results of interactive mixed effect model and autoregressive mixed effect model we come to the conclusion that autoregressive mixed effect model is superior to interactive mixed effect model.The index calculated by the autoregressive mixed effect model is more accurately and realistic.