【作者】 孔玲爽； 阳春华； 王雅琳； 桂卫华；

【Author】 Lingshuang Kong~(1,2),Chunhua Yang~1,Yalin Wang~1,Weihua Gui~1 1.School of Information Science and Engineering,Central South University,Changsha 410083,China 2.College of communication and Electric Engineering,Hunan University of Arts and Science,Changde 415000,China

【机构】 中南大学信息科学与工程学院； 湖南文理学院电气与信息工程学院；

【摘要】 将经典重构技术和粗糙集约简理论相结合,提出一种广义相空问重构方法,为多变量时间序列的预测构造有效的模型输入向量。首先,采用平均单步预测误差最小化方法确定嵌入维数,获取多变量时间序列的初始重构相空间;然后,对嵌入维数补偿一定裕量,构建时间序列决策表,并利用粗糙集约简理论删除冗余嵌入和冗余变量,实现广义相空间重构;最后,根据广义重构结果构造输入样本集,辨识预测模型参数。验证结果表明,提出的样本构造方法能使预测模型具有更好的泛化能力,具有一定的应用价值.更多还原

【Abstract】 In order to obtain the effective input vector for the prediction of multivariate time series,a generic phase space reconstruction method combining classical reconstruction technology with reduction theory of rough sets was proposed. Firstly,the embedding dimension was determined by minimizing the mean one-step prediction error and the original reconstruction phase space was obtained.Then,the original decision-table with redundant embedding dimensions for multivariate time series was set up and the RS reduction was used to delete the redundant dimensions and irrelevant variables and to reconstruct the generic phase space.Finally,the samples were extracted according to generic reconstruction results to identify the parameters of prediction model.Verification results show that the developed reconstruction method leads better generalization ability for the prediction model and it is feasible and worthwhile for application.更多还原