搜索

x

留言板

尊敬的读者、作者、审稿人, 关于本刊的投稿、审稿、编辑和出版的任何问题, 您可以本页添加留言。我们将尽快给您答复。谢谢您的支持!

姓名
邮箱
手机号码
标题
留言内容
验证码

基于鲁棒回声状态网络的混沌时间序列预测研究

李德才 韩敏

引用本文:
Citation:

基于鲁棒回声状态网络的混沌时间序列预测研究

李德才, 韩敏

Chaotic time series prediction based on robust echo state network

Li De-Cai, Han Min
PDF
导出引用
  • 针对回声状态网络模型易受异常点影响的问题, 提出一种基于拉普拉斯先验分布的鲁棒回声状态网络模型. 通过采用对异常点不敏感的拉普拉斯分布代替高斯分布作为模型输出的先验, 以增强网络对于异常点的抑制能力. 此外, 为解决由引入拉普拉斯分布所造成的求解困难的问题, 根据边际优化方法, 构建适当的替代函数, 使拉普拉斯先验等价转化为易于计算的高斯形式, 并通过贝叶斯方法实现模型参数的自适应估计. 仿真结果表明, 在异常点存在的情况下, 本文所提出的模型具有较好的鲁棒性, 并仍能保持较高的预测精度.
    Focusing on the problem that the echo state network is easily influenced by outliers, in this paper we propose a robust model based on the Laplace prior distribution. This is achieved by replacing the Gaussian distribution with the Laplace distribution as the prior of the model output, the Laplace prior is less sensitive to the outliers and can enhance the capbility of the model to restrict outliers. Furthermoer, to solve the problem arising from the introduction of the Laplace distribution, which makes the solving process of the method difficlut, the bound optimization algorithm is employed and a suitable surrogate function is established. Based on the bound optimization algorithm, the Laplace prior can be equivalently transformed into the form of Gaussian prior, which is easily computed, and it can also be use to estimate the model parameters adaptively. Simulation results illustrate that the proposed method can be robust when outliers exist, while remaining acceptable prediction accuracy.
    • 基金项目: 国家自然科学基金(批准号:61074096)资助的课题.
    [1]

    Chen S M, Hwang J R 2000 IEEE Trans. Systems, Man and Cybernetics-Part B 30 263

    [2]

    Dhanya C T, Kumar D N 2010 Advances in Warer Resources 33 327

    [3]
    [4]

    Du J, Cao Y J Liu Z J, Xu L Z, Jiang Q Y, Guo C X, Lu J G 2009 Acta Phys. Sin. 58 5997 (in Chinese) [杜 杰、 曹一家、 刘志坚、 徐立中、 江全元、 郭创新、 陆金桂 2009 物理学报 58 5997]

    [5]
    [6]
    [7]

    Leung H, Lo T, Wang S C 2001 IEEE Trans. Neural Network 12 1163

    [8]

    Cai J W, Hu S S, Tao H F 2007 Acta Phys. Sin. 56 6820 (in Chinese) [蔡俊伟、 胡寿松、 陶洪峰 2007 物理学报 56 6820]

    [9]
    [10]
    [11]

    Ma Q L, Zheng Q L, Peng H, Tan J W 2009 Acta Phys. Sin. 58 1410 (in Chinese) [马千里、 郑启伦、 彭 宏、 覃姜维 2009 物理学报 58 1410]

    [12]

    Farsa M A, Zolfaghari S 2010 Neurocomputing 73 2540

    [13]
    [14]

    Song Q S, Feng Z R, Li R H 2009 Acta Phys. Sin. 58 5057 (in Chinese) [宋青松、 冯祖仁、 李人厚 2009 物理学报 58 5057]

    [15]
    [16]
    [17]

    Jaeger H, Haas H 2004 Science 304 78

    [18]
    [19]

    Shi Z W, Han M 2007 IEEE Trans. Neural Networks 18 359

    [20]
    [21]

    Shi Z W, Han M 2007 Control and Decision 22 258 (in Chinese) [史志伟、 韩 敏 2007 控制与决策 22 258]

    [22]

    Han M, Mu D Y 2010 Control and Decision 25 531 (in Chinese) [韩 敏、 穆大芸 2010 控制与决策 25 531]

    [23]
    [24]
    [25]

    Ting J A, Dsouza A, Schaal S 2007 ICRA 2007 2489

    [26]
    [27]

    Zhong M J 2006 Neurocomputing 69 2351

    [28]

    Tipping M E 2001 Journal of Machine Learning Research 1 211

    [29]
    [30]

    Hong X, Chen S 2005 IEEE Trans. Systems, Man and Cybernetics-Part B 35 155

    [31]
    [32]

    Ding T, Zhou H C, Huang J H 2004 2010 Journal of Hydraulic Engineering 12 15 (in Chinese) [丁 涛、 周惠成、 黄健辉 2004 水力学报 12 15]

    [33]
    [34]
    [35]

    Islam M N, Sivakumar B 2002 Advances in Water Resources 25 179

    [36]
    [37]

    Fraser A M, Swinney H L 1986 Phys. Rev. A 33 1134

    [38]

    Kennel M B, Brown R, Abarbanel H D I 1992 Phys. Rev. A 45 3403

    [39]
  • [1]

    Chen S M, Hwang J R 2000 IEEE Trans. Systems, Man and Cybernetics-Part B 30 263

    [2]

    Dhanya C T, Kumar D N 2010 Advances in Warer Resources 33 327

    [3]
    [4]

    Du J, Cao Y J Liu Z J, Xu L Z, Jiang Q Y, Guo C X, Lu J G 2009 Acta Phys. Sin. 58 5997 (in Chinese) [杜 杰、 曹一家、 刘志坚、 徐立中、 江全元、 郭创新、 陆金桂 2009 物理学报 58 5997]

    [5]
    [6]
    [7]

    Leung H, Lo T, Wang S C 2001 IEEE Trans. Neural Network 12 1163

    [8]

    Cai J W, Hu S S, Tao H F 2007 Acta Phys. Sin. 56 6820 (in Chinese) [蔡俊伟、 胡寿松、 陶洪峰 2007 物理学报 56 6820]

    [9]
    [10]
    [11]

    Ma Q L, Zheng Q L, Peng H, Tan J W 2009 Acta Phys. Sin. 58 1410 (in Chinese) [马千里、 郑启伦、 彭 宏、 覃姜维 2009 物理学报 58 1410]

    [12]

    Farsa M A, Zolfaghari S 2010 Neurocomputing 73 2540

    [13]
    [14]

    Song Q S, Feng Z R, Li R H 2009 Acta Phys. Sin. 58 5057 (in Chinese) [宋青松、 冯祖仁、 李人厚 2009 物理学报 58 5057]

    [15]
    [16]
    [17]

    Jaeger H, Haas H 2004 Science 304 78

    [18]
    [19]

    Shi Z W, Han M 2007 IEEE Trans. Neural Networks 18 359

    [20]
    [21]

    Shi Z W, Han M 2007 Control and Decision 22 258 (in Chinese) [史志伟、 韩 敏 2007 控制与决策 22 258]

    [22]

    Han M, Mu D Y 2010 Control and Decision 25 531 (in Chinese) [韩 敏、 穆大芸 2010 控制与决策 25 531]

    [23]
    [24]
    [25]

    Ting J A, Dsouza A, Schaal S 2007 ICRA 2007 2489

    [26]
    [27]

    Zhong M J 2006 Neurocomputing 69 2351

    [28]

    Tipping M E 2001 Journal of Machine Learning Research 1 211

    [29]
    [30]

    Hong X, Chen S 2005 IEEE Trans. Systems, Man and Cybernetics-Part B 35 155

    [31]
    [32]

    Ding T, Zhou H C, Huang J H 2004 2010 Journal of Hydraulic Engineering 12 15 (in Chinese) [丁 涛、 周惠成、 黄健辉 2004 水力学报 12 15]

    [33]
    [34]
    [35]

    Islam M N, Sivakumar B 2002 Advances in Water Resources 25 179

    [36]
    [37]

    Fraser A M, Swinney H L 1986 Phys. Rev. A 33 1134

    [38]

    Kennel M B, Brown R, Abarbanel H D I 1992 Phys. Rev. A 45 3403

    [39]
计量
  • 文章访问数:  8463
  • PDF下载量:  688
  • 被引次数: 0
出版历程
  • 收稿日期:  2010-12-15
  • 修回日期:  2011-01-16
  • 刊出日期:  2011-05-05

/

返回文章
返回