基于RS-GA-BP神经网络的住宅房地产估价模型与应用

Residential Property Evaluation Model and Application Based on RS-GA-BP Neural Network

针对一般住宅房地产估价问题的非线性特征,融合粗糙集方法、遗传算法和神经网络算法的优势,提出了一种新的住宅房地产估价模型—基于粗糙集、遗传算法和BP算法集成的住宅房地产估价模型.首先对影响房地产估价因素进行离散化处理,其次利用粗糙集方法对估价因素进行约简,即精减BP神经网络的输入变量,最后利用遗传算法来优化BP神经网络初始权重和阈值.优化后的BP神经网络具有较好的处理非线性问题的能力,收敛速度和仿真精度较传统BP算法都有了明显的提高.选取某市工程案例进行实证分析,研究结果表明,新的估价方法能较客观准确地估测住宅房地产的价格,在住宅房地产估价中具备较高的实用性.

Considering the complexity of comprehensive evaluation of residential property, we integrate the advantages of rough set theory and genetic algorithm with neural network algorithm .And then we will bring up with a new model for evaluation of residential property, which is based on rough set theory, genetic algorithm and neural network algorithm. Firstly, we simplify the categorical attribute by rough set theory and optimize the input variable of neural network. Then, considering the drawbacks of traditional error counter propagation algorithm ,we make use of the overall searching ability of genetic algorithm and optimize the main parameters such as weights and threshold value of neural network through chromosome code.The new algorithm has not only the overall searching ability of genetic algorithm ,but also the learning ability and robustness of neural network.Compared to the traditional algorithm, the precision and speed of algorithm has been strengthened obviously at the same time.We choose one project case in Yiyang, Hunan to do empirical analysis.The research shows that neural network can rely on the features of project and assess the residential property objectively and properly. It suggests that this algorithm has great effectiveness in the evaluation of residential property.