点击次数：

论文类型：期刊论文

发表时间：2015-01-05

发表刊物：NEUROCOMPUTING

收录刊物：SCIE、EI

卷号：147

期号：1

页面范围：315-323

ISSN号：0925-2312

关键字：Echo state network; Multivariate time series; Online prediction; Square root cubature; Kalman filter

摘要：In this research, for online modeling and prediction of multivariate time series, we propose a novel approach termed squared root cubature Kalman filter-gamma echo state network (SCKF-gamma ESN). First, multivariate time series are modeled by using gamma echo state network (gamma ESN). Then, by using squared root cubature Kalman filter (SCKF), we update parameters of gamma ESN and predict future observations online. Furthermore, we add a robust outlier detection algorithm to SCKF to protect SCKF-gamma ESN from divergence caused by outliers. Finally, two numerical examples, by using a multivariate benchmark dataset and a real-world dataset, are conducted to substantiate the effectiveness and characteristics of the proposed SCKF-gamma ESN. (C) 2014 Elsevier B.V. All rights reserved.