基于BP神经网络误差校正的ARIMA组合预测模型

ARIMA Combined Prediction Model Based on BP Neural Network Error Correction

文章根据ARIMA时间序列模型和BP神经网络模型分别在处理线性空间预测问题和非线性空间预过测对问北题京中市的居优民势消,费建价立格了指一数种(C基PI于)B序P列神的经实网证络分误析差证校明正了的该差组分合自预回测归模移型动相平对均(于A单RI一MA预)测组模合型预在测C模PI型预,通测中的有效性,并利用该模型预测了未来一段时间北京市CPI的走势。

Based on the advantages of ARIMA time series model and BP neural network model in dealing with linear space prediction problems and nonlinear space prediction problems respectively, this paper establishes a differential autoregressive mov-ing average (ARIMA) combined prediction model based on BP neural network error correction. And then through the empirical analysis of the Beijing consumer price index (CPI) series, the paper proves the validity of the combined forecast model compared with the single forecast model in CPI forecast, and also uses this model to predict the CPI trend of Beijing for a period of time in the future.