The Modeling And Analysis On Early Warning For Real Estate Based On Neural Networks

Chinese real estate has experienced30years in development process so far, thedevelopment mode of which achieve to transform the planning system into the market system,but no matter what is in the stage of development and the operation mode, which always needthe regulate and control from the country and the government. China government implementsthe welfare system about the house in the planning system before1998. After the housingreform, house, as the main body of real estate, moves towards the market and develops veryquickly. It has very drastic fluctuations in the real estate market, which reflects in the pricefluctuations, much investment and speculative activities, and can directly or indirectly affectthe operation and stability of the national economy. It is particularly important for all the rolesto strengthen the regulation of the real estate market in the economic society.Against this background above, the thesis carries out the research about the systemmodeling and analysis for the real estate. Firstly, we read a lot of references, and know thecurrent research at home and abroad for the market regulation and early warning in the realestate. It introduce the regulatory system in the real estate market of several typical countriesand the early warning research to lay the foundation of finding the reasonable early warningscheme. Secondly, the thesis explains the real estate cycle theory, the real estate early warningtheory and elements and the real estate cycle factor theory, and then combines the principlesof economics and the real estate to lay the foundation of the theory.The thesis uses the way of early warning based on the Artificial Neural Networks. Firstly, itintroduces the preparatory works of early warning model, including screening index bystep-out time analysis and warning degree values defined interval division. Secondly, itselects the algorithm of the BP neural network, and explains the algorithm theory and variouslinks for the modeling and analysis application in detail, and achieves the combination ofthem, which does the original rational discourses for the empirical analysis.The thesis takes the real estate market in Tianjin for sample as an empirical analysis. Firstof all, it explains the basis for the Tianjin city as the sample cities and dividing the timeinterval, and then describes the course of development about the real estate industry in Tianjin.Secondly, it gets the detailed data about real estate early warning index in Tianjin by queryingStatistical Yearbook of Tianjin and the Tianjin Municipal Bureau of Statistics website toensure the authenticity and the accuracy in data. Finally, it needs to analyze and process thedata, and selects the early-warning indexes about the real estate in Tianjin. It uses the BPneural network modeling with the MATLAB software programming and ensures the trainingand parameter for the BP neural network. It also predicts the situation of the Tianjin real estatemarket in2012and comes to the conclusion of the degree of "hot" for the market operation.Finally, the thesis summarizes the results and conclusions, and emphatically analyzes shortage in the early warning system. It also gives us the prospects in the future research andoffers some policy proposals for the early warning research.