组合模型对居民消费价格指数序列的分析及预测

求和自回归移动平均模型(简称ARIMA)及支持向量回归模型(简称SVR)是两个重要且行之有效的分析及预测时间序列的方法.他们都能在一定程度上反映数据所包含的信息且信息不会完全重叠.为了能够各取所长,本文用这两种模型的组合模型对居民消费指数(CPI)进行了预测,结果显示组合模型提高了指数的预测精度.

The autoregressive integrated moving average model （ARIMA） and the support vector regression model （SVR） are two important and useful models for time series prediction.They can both reflects the information contained in the data and the information will never be completely overlap. In order to take the advantages of them both, this paper use the model of the combination of them to predic~ consumer price index （CPI）.The result shows that the combination model improve the prediction precision.