Design And Analysis The Early Warning System Of Real Estate Market In Xi'an City

According to the law of economics operation and the present situation of real estate in XI'AN city, to establish a scientific and effective early warning system is very important both for soundly developing the real estate in XI'AN city, and for efficiently utilising the resources. The early warning system is significant not only in theory but also in practice.On the basis of summarizing the progenitor's achievements, the early warning indices system of real estate market in XI'AN city is set up by analyzing the historical data, with the method of the principal component analysis and the fuzzy clustering analysis. And this indices system relies on the Economic Prosperity Theory and the Real Estate Warning System Theory. Then, the rationality of the indices system is verified by Granger causality test. Furthermore, A BP neural network is built to forecast the early warning indices data. Based on these data, the early warning model is designed and established. By using the model, the prosperity condition of the real estate market during the eleventh "five years plan" is forecasted, and the countermeasures are advised.The main innovations in this thesis are as follows:(1) Fifteen indices are selected from the twenty-seven original ones by combining the principal component analysis and the fuzzy clustering analysis, which compose the early warning indices system of real estate market in XI'AN city. And the calculation is executed by MATLAB software.(2) The Granger causality test method in economic analysis is used to examine the rationality of the early warning indices system, in order that its legitimacy can be inspected in the macroeconomic circumstance. And the calculation is performed by EVIEWS software.(3) First, the particular index forecasting method is established by BP neural network function. Then, the warning fields are classified by Prosperity Warning Theory. And the warning lines are determined with reference to the 3σmethod. As a result, the particular index monitoring and the synthetic indices monitoring and forecasting can be executed. So the early warning system of real estate market in XI'AN city is designed and set up. The calculation is implemented by BP neural network toolbox of MATLAB software. (4) In the process of building the early warning model, the powers of every index in the early warning indices system are decided by the ratios of the sub-numbers of the major component's characteristic vector. In this way, it can be avoided that the powers are unreasonably designed on average. And it is the base for reasonably monitoring and forecasting.(5) Forecast and analysis the real estate market during the eleventh "five years plan" in XI'AN city, and analysis the possible prosperities under different conditions. Moreover, some countermeasures are suggested.