多元混沌时间序列的因子回声状态网络预测模型被引量：19

Factor Echo State Network for Multivariate Chaotic Time Series Prediction

下载PDF

导出

摘要针对采用回声状态网络预测多元混沌时间序列时存在的病态解问题,本文建立了因子回声状态网络模型,通过因子分析(Factor analysis,FA)方法提取高维储备池状态矩阵的公因子,去除冗余和噪声成分.利用降维后的因子变量与期望输出之间的线性回归关系,求解网络未知参数.基于Lorenz序列和大连月平均气温–降雨量的仿真实验验证了本文所提模型的有效性. When an echo state network is used to predict mul-tivariate time series, there may exist ill-posed problem. This pa-per proposes a novel prediction model, named factor echo state network, to solve the problem. It uses a factor analysis （FA） al-gorithm to extract the common factors of the reservoir matrix, and to remove the redundancies and noises. Then the unknown output weights are calculated by linear regression of the output and common factors. The model is used to predict Lorenz series and monthly average temperature-rainfall time series in Dalian, and simulation results substantiate its effectiveness.

作者许美玲韩敏

机构地区大连理工大学电子信息与电气工程学部

出处《自动化学报》 EI CSCD 北大核心 2015年第5期1042-1046,共5页 Acta Automatica Sinica

基金国家自然科学基金(61374154) 国家重点基础研究发展计划(973计划)(2013CB430403)资助~~

关键词多元混沌时间序列预测回声状态网络因子分析 Multivariate chaotic time series, prediction, echostate network, factor analysis （FA）

分类号 O415.5 [理学—理论物理]

引文网络
相关文献

参考文献24

1Rong T Z, Xiao Z. Nonparametric interval prediction of chaotic time series and its application to climatic system. International Journal of Systems Science, 2013, 44(9): 1726-1732.
2韩敏,许美玲,任伟杰.多元混沌时间序列的相关状态机预测模型研究[J].自动化学报,2014,40(5):822-829. 被引量：13
3Inoussa G, Peng H, Wu J. Nonlinear time series modeling and prediction using functional weights wavelet neural network-based state-dependent AR model. Neurocomputing, 2012, 86(1): 59-74.
4Li P H, Li Y G, Xiong Q Y, Chai Y, Zhang Y. Application of a hybrid quantized Elman neural network in short-term load forecasting. International Journal of Electrical Power & Energy Systems, 2014, 55: 749-759.
5Yeh W C. New parameter-free simplified swarm optimization for artificial neural network training and its application in the prediction of time series. IEEE Transactions on Neural Networks and Learning Systems, 2013, 24(4): 661-665.
6玄兆燕,杨公训.经验模态分解法在大气时间序列预测中的应用[J].自动化学报,2008,34(1):97-101. 被引量：38
7Zeng Z G, Wang J. Improved conditions for global exponential stability of recurrent neural networks with time-varying delays. IEEE Transactions on Neural Networks, 2006, 17(3): 623-635.
8Zhang H G, Liu J H, Ma D Z, Wang Z S. Data-core-based fuzzy min-max neural network for pattern classification. IEEE Transactions on Neural Networks, 2011, 22(12): 2339-2352.
9Zhang H G, Liu D R, Luo Y H, Wang D. Adaptive Dynamic Programming for Control: Algorithms and Stability. London: Springer, 2013. 1-19.
10Jaeger H. The "Echo State" Approach to Analysing and Training Recurrent Neural Networks —— with An Erratum Note, GMD Report 148, German National Research Center for Information Technology, Germany, 2001.

二级参考文献73

1杨培才,周秀骥.气候系统的非平稳行为和预测理论[J].气象学报,2005,63(5):556-570. 被引量：41
2杨晓华,杨志峰,沈珍瑶,陆桂华,郦建强.水环境非线性时序预测的高精度RBF网络模型[J].水科学进展,2005,16(6):788-791. 被引量：9
3韩敏,史志伟,郭伟.储备池状态空间重构与混沌时间序列预测[J].物理学报,2007,56(1):43-50. 被引量：23
4张家树,党建亮,李恒超.时空混沌序列的局域支持向量机预测[J].物理学报,2007,56(1):67-77. 被引量：25
5Karunasinghea D S K, Liong S Y. Chaotic time series prediction with a global model: Artificial neural network [J]. J of Hydrology, 2006, 323(1-4): 92-105.
6Elman J L. Finding structure in time[J]. Cognitive Science, 1990, 14(2): 179-211.
7Takens F. Daynamical systems and turbulence, lecture notes in mathematics[M]. Berlin: Sringer, 1981.
8Peng D Z, Zhang Y. Dynamics of generalized PCA and MCA learning algorithms[J]. IEEE Trans on Neural Networks, 2007, 18(6): 1777-1784.
9Hartmann H, Becker S, King Lorenz. Predicting summer rainfall in the Yangtze river basin with neural networks[J]. Int J of Climatology, 2008, 28(7): 925- 936.
10Mika S, Scholkopf B, Smola A. Kernel PCA and denoising in feature space [C] . Advances in Neural Information Processing Systems. Cambridge: MTI Press, 1999: 524-536.