基于神经网络的城乡交错区土地估价方法

Neural networks based landprice evaluation method for urban/rural composition area

提出了城乡交错区土地价格评估的人工神经网络系统的基本构造,并在充分考虑城市规划对城乡交错区土地价格影响的基础上对神经网络的估价系统进行了设计.给出了构成城乡交错区土地价格的市场、区位、城市规划等方面影响因素的特性指标及特征指标的量化计算公式.以杭州市未来10年城市规划为基础,分析了杭州市城乡交错区各个片区的特征指标值,利用改进的BP神经网络模型对它们进行了训练和模拟,取得了良好的模拟结果,证明了ANN法用来预测城乡交错区地价的可行性和先进性.

The basic structure of price evaluation for urban/rural composition area based on the theory of artificial neural networks is put forward. The characteristic indexes such as market, location and urban planning which affect the price of urban/rural composition area are given and their calculations are also given. The authors take the Hangzhou's urban planning in the next ten years for example to analyze the characteristic indexes for special tracts, and use the improved BP Neural Networks to train and simulate the realistic price of urban/rural composition area. The simulation result shows that the ANN method for evaluating the price of urban/rural composition area is feasible and advanced in performance.