Research On The PSO-lSSVR Predication Model Of Urban Housing Prices

Along with the dramatic expansion of real estate industry in China, the residentialproduct as an important component in real estate industry has experienced greatincrease. The price of the residential product has been playing a lever role in economy,which also propels the industrialization and commercialization of residential products.In the meantime, the price has gained great amount of attention from the government,residence and real estate developers. Especially in recent years, the continuouslyincreasing price influences adversely the living quality of residence and the entire socialeconomy. Even though, the micro regulations from the government do work, theeffectiveness is not obvious. Based on present situation, this research aims to establish amodel to predict the price of urban residential products. According to the predictionmodel, the price trend can be forecast, the price of the residential products can beassessed more accurately, and thus, the development of real estate industry can beanalyzed, which is significant to ensure the healthy and stable development of realestate industry. In this research, the main contents are:①At first, after reviewing the domestic and foreign relevant researches of theurban housing prices’ predication, this dissertation proposes the methods to forecasthousing prices by using Least Squares Support Vector Machines (PSO-LSSVR Model)based on Particle swarm optimization. Least Squares Support can make up manyshortages of Artificial Neural Network Model and Support Vector Machine in theprocess of Modeling Data. Besides, Particle swarm optimization is able to optimizeparameters rapidly with a number of advantages like high accuracy and speediness.②Not only does it introduce the operating process of the PSO-LSSVRPredication Model of Urban Housing Prices by means of confirming the real estateindex system and criterion of classification, but also it presents the Framework of RealEstate Market Early Warning System based on introducing the PSO-LSSVR Predicationand Fuzzy and Grey Systems Theory, which helps the predication model to play its rolebetter.③It takes Beijing for example to develop the case analysis of the predicationmodel and assesses the health condition of the development of Beijing Real EstateMarket by building the corresponding PSO-LSSVR Predication Model of housingprices and trading volume. The case analysis demonstrates that least squares support vector regression (LSSVR) predication model based on Particle swarm optimization issuperior to the traditional support vector regression model. Meanwhile, it proves thevalidity of the PSO-LSSVR Predication Model to forecast the urban housing prices aswell.④It shows the implementation procedures of the whole Real Estate Market EarlyWarning System on the basis of the PSO-LSSVR Predication Model of housing pricesand trading volume and Real Estate Market Assessment model based on Fuzzy andGrey Systems. Moreover, it introduces the process of the system software realization.⑤Conclusion of theresults and the recommendation for future research.This research introduces the PSO-LSSVR Predication Model of Urban HousingPrices. Based on Fuzzy and Grey Systems, this research also indicates a set of methodsto predict and assess the real estate market. These two results are worthwhile forprediction and assessment of real estate markets. According to the results, this researchcan assist the administration department to carry out regulation policies, providemethods to ensure real estate industry develop healthily, constantly and stably, help thecorporates to make the investmentdecisions, and also complete the theory system forreal estate in China.