小波神经网络在房地产价格指数预测中的应用

Real Estate Price Indices Forcast by Using Wavelet Neural Network

随着房地产价格指数的作用充分显现,探求预测房地产价格指数的有效方法是需深入研究的方向。该文以中房上海住宅价格指数为例,首先对房地产价格指数序列性质进行分析,表明房地产价格指数是具有非线性特征的非平稳时间序列。采用小波神经网络对房地产价格指数进行预测,并将预测结果与指数平滑法和RBF神经网络预测做了对比。采用MATLAB对拟合和预测过程进行仿真。结果指标表明,在大样本数据的情况下,采用小波神经网络对房地产指数进行预测能够获得较好的效果。

As real estate price indices play more significant roles, it is necessary to give more effective method for forecasting real estate price indices, By exploiting the data from the Shanghai housing price index of China Real Estate Index System （CREIS）, this paper presents a wavelet neural network （WNN） model to give its forecasting. The comparisons among WNN, the exponential smoothing and RBF neural network, which are widely applicable, show that the forecasting of the WNN model is more effective given the large sample.