融GIS和BP神经网络的住宅房产评估模型

Evaluation model of residential property based on GIS and BP neural network model method

针对房产税征收对住宅房产价格批量评估的需求问题,该文提出融合GIS和BP神经网络的住宅房产批量评估模型。该模型以特征价格理论为依据,利用GIS数据分析方法构建了16个住宅房产价格的影响因子及量化指标,对房产交易案例库及待估房产进行批量量化,然后利用BP神经网络挖掘房产估价因子与价格之间的对应关系,对量化的待估房产进行批量评估,并开发了相应的软件系统。该文以赣州市章贡区10处待估房产为例,对模型进行验证。结果表明,估算出的房产价格与实际成交价格接近,模型具有较高准确性。

In view of the demand of real estate tax levy on the real estate price in bulk evaluation,this paper put forward the model of real estate quantity evaluation based on the fusion of GIS and BP（back propagation）neural network.Based on the characteristic price theory,16 residential real estate price factors and quantitative index were constructed by using GIS data analysis method,and the real estate transaction cases and the real estates were quantified in batches.Then combined with the strong nonlinear computing ability of BP neural network,the relationship between factors and the price of the real estate valuation could be excavated.Finally the corresponding software system was developed.Ten real estates in Zhanggong District of Ganzhou were taken as samples to verify the model.The results showed that the estimated real estate prices were close to the actual transaction price and the accuracy of this model was high.