The Research About Combined Forecasting Model Applied In House Prices In Chengdu

The real estate is an important part of the national economy, both by the overall level of national economic development constraints, but also played for the national economic development, the role of new economic growth point. Commercial housing price is the real estate market conditions, the most direct and most scientific reflection, The excessive growth of commercial housing prices have become a hot issue. If left unchecked, will affect the healthy development of the real estate market continues to affect the macroeconomic and social stability of. Commercial housing price is closely related with people's lives, whether from the perspective of development of national economies, or from meeting the basic needs of the people's point of view, commercial housing price changes and trends are essential. Therefore, it is necessary to forecast commercial house price in a scientific method, Not only to ensure our government develop the real estate market reasonably which is conducive to healthy and stable development of policiesthe, And let every real estate developers take market as the guide, Reasonable investment in new property development and establish reasonable sales prices,to ensure their own interests; Or the wish of each ordinary home buyers. But because the price of the influence factors of the house is made of economics is a complex problem prediction.This paper selects chengdu commercial data from 1999 to 2008(Source: web site in chengdu bureau of statistics yearbook).Using three kinds of forecasting methods gray GM(1,1), three -times exponential smoothing and unary linear regression respectively, established three different commercial housing price of chengdu forecasting model, Analysis of the model prediction error. It is clear that the price of commercial housing in Chengdu forecast, Grey GM (1,1) is better than the other two forecasting methods, the lower error, the unary linear regression prediction error is higher. If you think that a prediction method of forecasting error, put this kind of forecasting abandoned, this may result in loss of some useful information, a single forecast accuracy at a low point of the forecast. In order to build an effective data, the introduction of induced ordered weighted averaging (IOWA) operator, establish induction orderly a weighted average of combination forecast model. Select three error index :square error(SSE),Mean-square error(MSE),Mean-square error percentage(MSPE), calculated three single forecasting model of chengdu housing prices average prediction error and the weighted average induction and orderly combined forecasting error, One grey model GM (1,1), SSE=174180,MSE=41.7346,MSPE=0.0206 ; three -times exponential smoothing model SSE=325910 ,MSE=57.0889,MSPE=0.0209;unary linear regression model SSE=414440,MSE=64.3772 , MSPE=0.0218 ; IOWA operator combination forecast model SSE=15686,MSE=12.5245, MSPE=0.0045. Combination of induced ordered weighted average of the three forecast error indicator values are significantly lower than the prediction model of three individual Chengdu index value of commercial housing price forecast error, show that the induced ordered weighted average combination forecasting model can effectively improve the prediction accuracy. Therefore, the applicability of the model is more powerful. According to IOWA combination forecasting model based calculated average price of commercial housing in 2010 was 6469.4 yuan in Chengdu.