Comparative Study Of Real Estate Price Evaluation Model In Tianjin City

In recent years, the reform of real estate transaction tax have launched in many cities, and the research of the technology of mass appraisal in real estate appraisal has obtained some result. However, research on evaluation of domestic price have generally focused on the linear regression model.The linear pricing model for asset appraisal often is poor accuracy, and hardly using in large scale data with efficient processing. Foreign scholars put forward a series of combination of the Chaos Theory, Grey Mathematics and Artificial Intelligence for nonlinear model.But in China, research for this kind of model didn’t attract the academic attention., and the study about Artificial Neural Network is more scarce. ANN model is the development and application of computer artificial intelligence technology.ANN has a well ability of learning because of its structure that is a kind of simulation model on human neurons.So the ANN has the excellent processing capability for the relationship between the variables of uncertain nonlinear. The neural network model was originally used in the decision analysis and medical diagnosis and so on,until the end of the last century it began to be used in the field of economy and finance, and continued to develop and mature.This paper compared the traditional tools with the Hedonic theory, and testify artificial neural network model is good at nonlinearity analysis. The multivariate linear regression and BP neural network is used to model the characteristic price theory pricing models are established in the course of the study, and by the Mean Absolute Percentage Error and Hit Rate and other indicators for the comprehensive evaluation of the two models. Finally, found that the artificial neural network model, rather than the linear model,has the advantages on pricing forecast,and its measurement method is more suitable for large-scale batch pricing.This study contributes to the real estate market participants and the relevant government departments to establish efficient and accurate real estate volume pricing model.