EEMD和布谷鸟搜索算法优化SVR的混沌时间序列预测  

Chaotic time series prediction based on SVR optimized by EEMD and cuckoo search algorithm

作　　者：乐洋[1,2] 江畅 陈德良 LE Yang;JIANG Chang;CHEN Deliang(Nanjing University of Posts and Telecommunications,Nanjing 210023,China;Smart Health Big Data Analysis and Location Services Engineering Lab of Jiangsu Province,Nanjing 210023,China)

机构地区：[1]南京邮电大学,江苏南京210023 [2]江苏省智慧健康大数据分析与位置服务工程实验室,江苏南京210023

出　　处：《现代电子技术》2022年第15期118-122,共5页Modern Electronics Technique

基　　金：江苏省智慧健康大数据分析与位置服务工程实验室开放研究基金(SHEL219002)。

摘　　要：为了对混沌时间序列的预测精度进行提升,提出组合预测模型,它对支持向量回归(SVR)、布谷鸟搜索(CS)与经验模态分解(EEMD)进行了综合。首先对混沌时间序列通过集合经验模态分解为一组去除噪声且相对稳定的子序列;借助CS算法对SVR参量加以优化,进而构建以SVR为基础的预测模型,由此获取原始序列的预测大小;接着以太阳黑子混沌时间序列为对象,对其进行预测实验,并与SVR、CS⁃SVR和EEMD⁃SVR的预测性能进行比较。结果显示,通过CS算法进行优化后,能够让SVR具有更快的收敛速度,使之预测精度有了明显的提升,同时也提升了它的泛化能力。A combined prediction model is proposed to improve the prediction accuracy of chaotic time series,in which the support vector regression(SVR),the cuckoo search(CS)algorithm and the ensemble empirical mode decomposition(EEMD)are synthesized.The chaotic time series is decomposed into a group of relatively stable subsequences with noise being removed by the ensemble empirical mode.The SVR parameters are optimized with the CS algorithm to construct a prediction model based on SVR,so as to obtain the predictive size of the original sequence.And then,the prediction experiment was carried out by taking the sunspot chaotic time series as the object.The predictive performance of the model are contrasted with those of SVR,CS⁃SVR and EEMD⁃SVR.The results show that the SVR optimized by the CS algorithm have a faster convergence speed,so that its prediction accuracy is improved obviously and its generalization ability is also improved.

关 键 词：混沌时间序列预测 集合经验模态分解 布谷鸟搜索算法 支持向量回归 太阳黑子 回归拟合 

分 类 号：TN911.1-34[电子电信—通信与信息系统]