论文部分内容阅读
将小波分析和支持向量回归(SVR)模型引入消费者物价指数CPI的时间序列分析中,利用小波降噪对原始时间序列进行小波变换,充分提取和分离金融时间序列的各种隐周期和非线性,把小波分解序列的特性和分解数据随尺度倍增而倍减的规律充分用于SVR支持向量回归模型的建模。将该方法应用于中国宏观经济指标CPI的分析与预测,可以有效预测CPI的变动方向,并显著提高CPI的预测精度。
The wavelet analysis and support vector regression (SVR) model are introduced into the time series analysis of CPI. The wavelet transform is used to transform the original time series to fully extract and separate the various implicit periods and nonlinearities of the financial time series , The law of multiplication and multiplication of the characteristics and decomposition data of the wavelet decomposition sequence with the scale is fully used in the modeling of the SVR support vector regression model. Applying this method to the analysis and forecasting of China’s macroeconomic indicators CPI can effectively predict the direction of the CPI’s change and significantly improve the prediction accuracy of the CPI.