基于BP神经网络与多元回归的公路客运量预测研究

Highway Passenger Volume Prediction Base on Multiple Regression and BP Neural Network

根据公路客运量的历史数据及相关影响因素数据,采用SPSS及MATLAB软件分别建立多元回归及BP神经网络预测模型,通过将2种模型输出的预测结果与实际数据进行对比,得出采用多元回归方法在预测客运量变化过程中所产生的误差远小于BP神经网络方法所产生的误差,认为,当数据样本量较小时,多元线性回归预测模型优于BP神经网络预测模型。

SPSS and MATLAB software has been used to establish the BP Neural Network model and Multiple Linear Regression model respectively,based on the historical data of highway passenger transport volumeand the related influencing factors. Though analyzing and comparing the result of BP Neural Network and Multiple Linear Regression result with actual value,it is found that the error caused by Multiple Linear Regression method in forecasting the change of passenger volume is much less than that of BP Neural Network method.Therefore,if the data base is tiny,the Multiple Linear Regression forecast model is superior to the BP Neural Network model.