A GA-BP Neural Network Based On Gray Correlation For Residential Residential Tax Base Batch Evaluation

In the past China property tax has been based on the operation of real estate tax,but in order to control the real estate market,expand the local government revenue sources,in January 28,2011,Chongqing and Shanghai have began the implementation of a property tax,and have met the conditions of the individual housing zoned into the tax range to the real estate tax base value.How to evaluate the estate tax base value become the focus of attention,accurate assessment of the real estate is the basic premise of property tax,real estate tax assessment reflects its important position in the whole reform process.Once the real estate tax popularize in the country,real estate tax involves the scope and number of dramas,for specific the previous assessment completed by hand in traditional real estate cannot meet the current assessment requirements,therefore,how to improve the scientific nature of real estate tax assessmen scientific and accuracy,ensure efficiency and low cost to complete the work of tax base valuation has become the real estate valuation industry an important topic that need to study and solve.Confronted with the real estate tax assessment has many objects of the background,mass appraisal of real estate tax evaluation is imperative,so mass appraisal is the best scheme to solve this problem.This article based on the theory of real estate tax assessment on the value of the type of residential real estate tax assessment,pointed out the current domestic traditional market method,cost method and income method can not meet the requirements of the tax assessment,the real estate tax as the evaluation of the mass appraisal,and analyzes the characteristics of mass appraisal method for the problem of the mass appraisal of ordinary residential tax base,integration of grey relational theory,GA algorithm and the advantages of BP neural network model put forward based on the grey incidence analysis theory of mass appraisal of a new residential tax base residents,the mass appraisal model of GA algorithm and BP neural network integrated residential tax base.The frist and foremost,the main factors selected effect of residential real estate prices through the gray relational analysis method,which simplifies the BP neural network input variables;Additionally,according to the BP neural network model's disadvantage,and The assessment of the tax base valuation using genetic algorithms to optimize the BP neural network of the initial value,based on the grey relational analysis of GA-BP neural network evaluation model.Finally,this paper selects the Nanchangfor empirical analysis,respectively,the establishment of traditional BP neural network model and GA-BP neural network model,after comparative analysis,we can draw a conclusion that the GA algorithm to optimize the BP neural network not only improve the traditional BP algorithm for nonlinear processing ability,but also improve the traditional BP algorithm's convergence speed and generalization ability and prediction results are more close to the real value.The experimental results show that the new method can fast and accurate estimation of residential real estate prices in the residential real estate has good application prospects.