Research Of Real Estate Appraisal Method Based On Genetic Neural Network

With the blossom of real estate market and the increase of marketing quantity, the industry of real estate appraisal has developed rapidly. But at present both study of real estate appraisal theory and development of the industry are not well-rounded. Real estate appraisal approach depends too much on experience of appraisal personnel, and is lack of accurate appraisal basis, hard to guarantee the quality of the appraisal result. These problems badly handicap the good development of the real estate appraisal industry.Considering the complicated and nonlinear relations between the real estate price and its influence factors, artificial neural network is used as real estate appraisal technology approach in this paper. Artificial neural network is used to confirm the mapping relation between the real estate price and its influence factors, so that the deviation of appraisal result due to the subjective factor of the appraisal personnel can be effectively reduced. Considering the defects of the slow convergence rate and falling into the local optimal solution easily of BP algorithm, genetic algorithm is used to optimize the weight and bias values of BP neural network, and make the real estate appraisal run at a more reasonable network structure. A real estate appraisal platform based on genetic neural network is constructed, and the feasibility of real estate appraisal based on genetic neural network is verified with Matlab 7.0.Considering the two important influence factors of real estate price——real estateoccupancy rate and rent level can't be assured accurately, a solution that data warehouse technology is used to real estate appraisal is proposed in this paper. The structure of the real estate appraisal oriented data warehouse is given, and then the multidimensional dataset that is used to OLAP analysis is set up. So that the appraisal personnel is able to get information of real estate occupancy rate and rent level of appraised object rapidly, flexibly, directly. And then, a real estate data warehouse system prototype is designed and realized with VC++6.0 and Microsoft SQL Server Analysis Services 2000.