The Real Estate Warning Study Based On Multivariate Logistic Regression Model

With the development of national economy, the real estate industry has becomethe important industry to promote economic growth, because of its high industriesassociation and strong driving. In recent years, the real estate market developsrapidly, but there are many problems, such as: housing price increasing rapidly,housing vacant volume high, and the obvious contradiction between housing supplyand demand. In order to promote the real estate market steadily and developmenthealthy, to avoid the volatility of the real estate market, it is necessary to establish ascientific, practical real estate market early warning system.The research ideas of this paper are as follows:The main content of the first part are background and significance of theresearch, the contents, framework and methods of the research and finally putsforward the innovation of this paper: it is the first time that the multivariate logisticregression model applied to the real estate market early warning in city of Jinan.The second part is status summaries of the real estate market. The maincontents include the basic theory of the early warning systems, the comparativeanalysis of the early warning methods, the early warning indicator system of the realestate market and the practical results of the early warning system in the real estatemarket in our country. The warning methods of the real estate market include theblack warning method, yellow warning method, white warning method, green andred warning method. In this paper, we choose the Model Prediction Method bycomparative analysis of working principle and basic characteristics of the variousearly warning methods.In the third part, we choose a suitable warning method of the real estate marketthrough the contrast analysis. First we choose the linear regression model which isthe most representative of in the Model Prediction Method.Second we analyze thelogistic regression model. And last it shows the superiority of the logistic regressionmodel in the real estate market in early warning from selection of variables, solving qualitative problem, and the fitting of the things in the development process and thefinal results in the form.The fourth part is to construct the index system of the real estate market. Themain contents include the choice of warning index, the determination of the warningsigns indicators and the division of the warning interval. We choose the real estateprice index as warning index. The warning signs indicators will be divided into threesections: the coordination between the real estate industry and national economy, thedevelopment speed and internal balance of the real estate industry. Each section isincluded of several specific warning signs index. Using3σ method, the warninginterval classified as cold, cool, normal, partial heat and overheating of five statesand representative with blue, light blue, green, yellow and red respectively.The fifth part is the empirical analysis, taking the city of Jinan as example. Wetake the real estate market-related quota of the city of Jinan from1996to2010asbasic data. By using linear regression and logistic regression respectively, we dowarning research for the city of Jinan, get the different results and put forward policyrecommendations.The last part of this paper is the conclusion and prospect. We summarize themain conclusion of the research, at the same time, point out the deficiency andimprovement.