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

一种识别关联维数无标度区间的新方法

周双 冯勇 吴文渊

引用本文:
Citation:

一种识别关联维数无标度区间的新方法

周双, 冯勇, 吴文渊

A novel method to identify the scaling region of correlation dimension

Zhou Shuang, Feng Yong, Wu Wen-Yuan
PDF
导出引用
  • 在计算关联维数过程中, 为了减少人为因素识别无标度区间带来的误差, 提出一种基于模拟退火遗传模糊C均值聚类识别无标度区间的新方法. 该方法根据无标度区间对应曲线的二阶导数在零附近波动的变化特征, 利用分类算法进行识别. 首先对双对数关联积分的离散数据进行二阶差分; 然后利用模拟退火遗传模糊C均值聚类方法对该数据进行分类, 选出在零附近波动的数据; 再剔除粗大误差保留有效数据; 最后进行统计分析识别出线性度最好的作为无标度区间. 应用新方法对两个著名的混沌系统Lorenz 和Henon 进行了仿真, 计算结果与理论值非常符合. 实验表明, 所提出的新方法与主观识别、K-means和2-means方法比较, 可以有效自动识别无标度区间, 减少误差, 计算结果更加精确.
    A random fractal exhibits self-similarity over the scaling region, this is different from the regular fractal. The scaling region obtained by the proper method for the exact fractal dimension is very important. And the correlation dimension is one of the fractal dimensions which is used widely in many fields. Therefore, it is necessary and timely to identify the scaling region that plays a critical role in calculating the correlation dimension accurately in various chaotic systems. Visual identification is widely used to determine the scaling region as a quick and simple subjective method. However, this method may lead to an inaccurate result in Grassberger Procaccia algorithm. In order to reduce the error caused by human factors from computing the correlation dimension, a novel method of identifying the scaling region based on simulated annealing genetic fuzzy C-means clustering algorithm is proposed. This new method is based on the fluctuating characteristics that the second-order derivative of the curve within the scaling region is zero or nearly zero. Firstly, the second-order differential of the double logarithm correlation integral discrete data is calculated. Secondly, the simulated annealing genetic fuzzy C-means clustering method is used for dividing the data into three groups: positive fluctuation data, zero fluctuation data, and negative fluctuation data. The zero fluctuation data are selected to retain, the rest is excluded. Thirdly, the 3 σ criteria are used for excluding gross errors to retain those valid from the zero fluctuation data. Fourthly, the data of the consecutive nature point interval are chosen from the retained data. Finally, the regression analysis and statistical test are used to identify the scaling region from the data valid. In order to verify the effectiveness of the proposed method, it is used to simulate the Lorenz and Henon systems. The calculated results are in good agreement with the theoretical values. Experimental results show that the proposed new approach is easy to operate, more efficient and more accurate than the subjective recognition, K-means method, and 2-means method in identifying the scaling region.
    • 基金项目: 国家自然科学基金(批准号:11301524)、重庆市科技攻关项目(批准号:cstc2012ggB40004)和中国科学院西部之光联合学者资助的课题.
    • Funds: Project supported by the National Natural Science Foundation of China (Grant No. 11301524), the Chongqing Science and Technology Key Project, China (Grant No. cstc2012ggB40004), and the CAS western light program.
    [1]

    Ostryakov V M, Usoskin I G 1990 Solar Physics 127 405

    [2]

    Wang W J, Wu Z T 2000 Journal of Shanghai Jiaotong 34 1265 (in Chinese) [汪慰军, 吴昭同 2000 上海交通大学学报 34 1265]

    [3]

    Casaleggio A, Corana A 2000 Chaos, Solitons & Fractals 11 2017

    [4]

    Huang G R, Rui X F 2004 Advances in Water Science 15 255 (in Chinese) [黄国如, 芮孝芳 2004 水科学进展 15 255]

    [5]

    Huang R S, Huang H 2007 Chaos and Its Applications (Second Edition) (Wuhan: Wuhan University Press) p217 (in Chinese) [黄润生, 黄浩 2007 混沌及其应用 (第二版) (武汉: 武汉大学出版社) 第217页]

    [6]

    Sheng Y G, Xu Y, Li Z S, Wu D, Sun Y H, Wu Z H 2005 Acta Phys. Sin. 54 221 (in Chinese) [盛永刚, 徐耀, 李志宏, 吴东, 孙予罕, 吴中华 2005 物理学报 54 221]

    [7]

    Yokoya N, Yamamoto K, Funakubo N 1989 Computer Vision, Graphics, and Image Processing 46 284

    [8]

    Maragos P, Sun F K 1993 IEEE Transactions on signal Processing 41 108

    [9]

    Kim H S, Eykholt R, Salas J D 1999 Physica D 127 48

    [10]

    Harikrishnan K P, Misra R, Ambika G, Kembhavi A K 2006 Physica D 215 137

    [11]

    Wang F Q, Luo C S, Cheng G X 1993 Chinese Journal of Computational Physics 10 345 (in Chinese) [汪富泉,罗朝盛,陈国先 1993 计算物理 10 345]

    [12]

    Bolea J, Laguna P, Remartínez J M, Rovira E, Navarro A, Bailón R 2014 Computational and Mathematical Methods in Medicine 2014 1

    [13]

    Judd K 1994 Physica D 71 421

    [14]

    Xiong J, Chen S K, Wei W, Liu S, Guan W 2014 Acta Phys. Sin. 63 200504 (in Chinese) [熊杰, 陈绍宽, 韦伟, 刘爽, 关伟 2014 物理学报 63 200504]

    [15]

    Dang J W, Shi Y, Huang J G 2003 Computer Engineering and Applications 23 35 (in Chinese) [党建武, 施怡, 黄建国 2003 计算机工程与应用 23 35]

    [16]

    Wu Z C 2002 Acta Geodaetica et Cartographica Sinica 31 240 (in Chinese) [巫兆聪 2002 测绘学报 31 240]

    [17]

    Du B Q, Jia Z W, Tang G J 2013 Journal of Vibration and Shock 32 40 (in Chinese) [杜必强, 贾子文, 唐贵基 2013 震动与冲击 32 40]

    [18]

    Wang C D, Ling D, Miao Q 2012 Computer Engineering and Applications 48 9 (in Chinese) [王成栋, 凌丹, 苗强 2012 计算机工程与应用 48 9]

    [19]

    Wu H S, Ni L P, Zhang F M, Zhou X, Du J Y 2014 Control and Decision 29 455 (in Chinese) [吴虎胜, 倪丽萍, 张凤鸣, 周漩, 杜继勇 2014 控制与决策 29 455]

    [20]

    Yang H Y, Ye H, Wang G Z, Pan G D 2008 3rd IEEE Conference on Industrial Electronics and Applications Singapore, Jun 3-5, 2008 p1383

    [21]

    Ji C C, Zhu H, Jiang W 2010 Chinese Science Bulletin 31 3069 (in Chinese) [姬翠翠, 朱华, 江炜 2010 科学通报 31 3069]

    [22]

    Wang A L, Yang C X 2002 Acta Phys. Sin. 51 2719 (in Chinese) [王安良, 杨春信 2002 物理学报 51 2719]

    [23]

    Grassberger P, Procaccia I 1983 Physical Review Letters 50 346

    [24]

    Grassberger P, Procaccia I 1983 Physica D 9 189

    [25]

    Kirkpatrick S, Gelatt J C D, Vecchi M P 1983 Science 220 671

    [26]

    Holland J H 1975 Adaptation in Natural and Artifical Systems (Ann Arbor: The University of Michigan Press)

    [27]

    Jain A K 2010 Pattern Recognition Letters 31 651

    [28]

    Bai L Y, Hu S Y, Liu S H 2005 Computer Engineering and Applications 41 56 (in Chinese) [白莉媛, 胡声艳, 刘素华 2005 计算机工程与应用 41 56]

    [29]

    Liang J W, Chen L C, He G 2001 Error Theory and Data Processing (Revised Edition) (Beijing: China Metrology Press) p57 (in Chinese) [梁晋文, 陈林才, 何贡 2001 误差理论与数据处理 (修订版) (北京: 中国计量出版社) 第57页]

  • [1]

    Ostryakov V M, Usoskin I G 1990 Solar Physics 127 405

    [2]

    Wang W J, Wu Z T 2000 Journal of Shanghai Jiaotong 34 1265 (in Chinese) [汪慰军, 吴昭同 2000 上海交通大学学报 34 1265]

    [3]

    Casaleggio A, Corana A 2000 Chaos, Solitons & Fractals 11 2017

    [4]

    Huang G R, Rui X F 2004 Advances in Water Science 15 255 (in Chinese) [黄国如, 芮孝芳 2004 水科学进展 15 255]

    [5]

    Huang R S, Huang H 2007 Chaos and Its Applications (Second Edition) (Wuhan: Wuhan University Press) p217 (in Chinese) [黄润生, 黄浩 2007 混沌及其应用 (第二版) (武汉: 武汉大学出版社) 第217页]

    [6]

    Sheng Y G, Xu Y, Li Z S, Wu D, Sun Y H, Wu Z H 2005 Acta Phys. Sin. 54 221 (in Chinese) [盛永刚, 徐耀, 李志宏, 吴东, 孙予罕, 吴中华 2005 物理学报 54 221]

    [7]

    Yokoya N, Yamamoto K, Funakubo N 1989 Computer Vision, Graphics, and Image Processing 46 284

    [8]

    Maragos P, Sun F K 1993 IEEE Transactions on signal Processing 41 108

    [9]

    Kim H S, Eykholt R, Salas J D 1999 Physica D 127 48

    [10]

    Harikrishnan K P, Misra R, Ambika G, Kembhavi A K 2006 Physica D 215 137

    [11]

    Wang F Q, Luo C S, Cheng G X 1993 Chinese Journal of Computational Physics 10 345 (in Chinese) [汪富泉,罗朝盛,陈国先 1993 计算物理 10 345]

    [12]

    Bolea J, Laguna P, Remartínez J M, Rovira E, Navarro A, Bailón R 2014 Computational and Mathematical Methods in Medicine 2014 1

    [13]

    Judd K 1994 Physica D 71 421

    [14]

    Xiong J, Chen S K, Wei W, Liu S, Guan W 2014 Acta Phys. Sin. 63 200504 (in Chinese) [熊杰, 陈绍宽, 韦伟, 刘爽, 关伟 2014 物理学报 63 200504]

    [15]

    Dang J W, Shi Y, Huang J G 2003 Computer Engineering and Applications 23 35 (in Chinese) [党建武, 施怡, 黄建国 2003 计算机工程与应用 23 35]

    [16]

    Wu Z C 2002 Acta Geodaetica et Cartographica Sinica 31 240 (in Chinese) [巫兆聪 2002 测绘学报 31 240]

    [17]

    Du B Q, Jia Z W, Tang G J 2013 Journal of Vibration and Shock 32 40 (in Chinese) [杜必强, 贾子文, 唐贵基 2013 震动与冲击 32 40]

    [18]

    Wang C D, Ling D, Miao Q 2012 Computer Engineering and Applications 48 9 (in Chinese) [王成栋, 凌丹, 苗强 2012 计算机工程与应用 48 9]

    [19]

    Wu H S, Ni L P, Zhang F M, Zhou X, Du J Y 2014 Control and Decision 29 455 (in Chinese) [吴虎胜, 倪丽萍, 张凤鸣, 周漩, 杜继勇 2014 控制与决策 29 455]

    [20]

    Yang H Y, Ye H, Wang G Z, Pan G D 2008 3rd IEEE Conference on Industrial Electronics and Applications Singapore, Jun 3-5, 2008 p1383

    [21]

    Ji C C, Zhu H, Jiang W 2010 Chinese Science Bulletin 31 3069 (in Chinese) [姬翠翠, 朱华, 江炜 2010 科学通报 31 3069]

    [22]

    Wang A L, Yang C X 2002 Acta Phys. Sin. 51 2719 (in Chinese) [王安良, 杨春信 2002 物理学报 51 2719]

    [23]

    Grassberger P, Procaccia I 1983 Physical Review Letters 50 346

    [24]

    Grassberger P, Procaccia I 1983 Physica D 9 189

    [25]

    Kirkpatrick S, Gelatt J C D, Vecchi M P 1983 Science 220 671

    [26]

    Holland J H 1975 Adaptation in Natural and Artifical Systems (Ann Arbor: The University of Michigan Press)

    [27]

    Jain A K 2010 Pattern Recognition Letters 31 651

    [28]

    Bai L Y, Hu S Y, Liu S H 2005 Computer Engineering and Applications 41 56 (in Chinese) [白莉媛, 胡声艳, 刘素华 2005 计算机工程与应用 41 56]

    [29]

    Liang J W, Chen L C, He G 2001 Error Theory and Data Processing (Revised Edition) (Beijing: China Metrology Press) p57 (in Chinese) [梁晋文, 陈林才, 何贡 2001 误差理论与数据处理 (修订版) (北京: 中国计量出版社) 第57页]

  • [1] 赵大帅, 孙志, 孙兴, 孙怀得, 韩柏. 基于分形理论的微间隙空气放电. 物理学报, 2021, 70(20): 205207. doi: 10.7498/aps.70.20210362
    [2] 周双, 冯勇, 吴文渊. 太阳高纬和低纬活动现象的混沌与分形特征. 物理学报, 2015, 64(24): 249601. doi: 10.7498/aps.64.249601
    [3] 熊杰, 陈绍宽, 韦伟, 刘爽, 关伟. 基于多重分形去趋势波动分析法的交通流多重分形无标度区间自动识别方法. 物理学报, 2014, 63(20): 200504. doi: 10.7498/aps.63.200504
    [4] 蔡建超, 郭士礼, 游利军, 胡祥云. 裂缝-孔隙型双重介质油藏渗吸机理的分形分析. 物理学报, 2013, 62(1): 014701. doi: 10.7498/aps.62.014701
    [5] 尚慧琳. 受迫Holmes-Duffing系统安全域分形及时滞速度反馈控制. 物理学报, 2012, 61(18): 180506. doi: 10.7498/aps.61.180506
    [6] 行鸿彦, 龚平, 徐伟. 海杂波背景下小目标检测的分形方法. 物理学报, 2012, 61(16): 160504. doi: 10.7498/aps.61.160504
    [7] 杨娟, 卞保民, 闫振纲, 王春勇, 李振华. 典型随机信号特征参数统计分布的分形特性. 物理学报, 2011, 60(10): 100506. doi: 10.7498/aps.60.100506
    [8] 杨娟, 卞保民, 彭刚, 李振华. 随机信号双参数脉冲模型的分形特征. 物理学报, 2011, 60(1): 010508. doi: 10.7498/aps.60.010508
    [9] 刘耀民, 刘中良, 黄玲艳. 分形理论结合相变动力学的冷表面结霜过程模拟. 物理学报, 2010, 59(11): 7991-7997. doi: 10.7498/aps.59.7991
    [10] 张丽, 刘树堂. 薄板热扩散分形生长的环境干扰控制. 物理学报, 2010, 59(11): 7708-7712. doi: 10.7498/aps.59.7708
    [11] 姜泽辉, 赵海发, 郑瑞华. 完全非弹性蹦球倍周期运动的分形特征. 物理学报, 2009, 58(11): 7579-7583. doi: 10.7498/aps.58.7579
    [12] 张程宾, 陈永平, 施明恒, 付盼盼, 吴嘉峰. 表面粗糙度的分形特征及其对微通道内层流流动的影响. 物理学报, 2009, 58(10): 7050-7056. doi: 10.7498/aps.58.7050
    [13] 孟田华, 赵国忠, 张存林. 亚波长分形结构太赫兹透射增强的机理研究. 物理学报, 2008, 57(6): 3846-3852. doi: 10.7498/aps.57.3846
    [14] 李 彤, 商朋见. 多重分形在掌纹识别中的研究. 物理学报, 2007, 56(8): 4393-4400. doi: 10.7498/aps.56.4393
    [15] 疏学明, 方 俊, 申世飞, 刘勇进, 袁宏永, 范维澄. 火灾烟雾颗粒凝并分形特性研究. 物理学报, 2006, 55(9): 4466-4471. doi: 10.7498/aps.55.4466
    [16] 高国良, 钱昌吉, 李 洪, 谷温静, 黄晓虹, 叶高翔. 无格点基底表面的杂质分布对分枝状凝聚体的影响. 物理学报, 2006, 55(7): 3349-3354. doi: 10.7498/aps.55.3349
    [17] 刘海峰, 代正华, 陈峰, 龚欣, 于遵宏. 混沌动力系统小波变换模数的关联维数. 物理学报, 2002, 51(6): 1186-1192. doi: 10.7498/aps.51.1186
    [18] 王安良, 杨春信. 评价奇怪吸引子分形特征的Grassberger-Procaccia算法. 物理学报, 2002, 51(12): 2719-2729. doi: 10.7498/aps.51.2719
    [19] 郭立新, 吴振森. 二维分数布朗运动(FBM)随机粗糙面电磁散射的基尔霍夫近似. 物理学报, 2001, 50(1): 42-47. doi: 10.7498/aps.50.42
    [20] 李 岩, 何大韧. 一类低反射率分形膜系的设计. 物理学报, 2000, 49(11): 2171-2175. doi: 10.7498/aps.49.2171
计量
  • 文章访问数:  4226
  • PDF下载量:  279
  • 被引次数: 0
出版历程
  • 收稿日期:  2014-12-10
  • 修回日期:  2015-02-09
  • 刊出日期:  2015-07-05

一种识别关联维数无标度区间的新方法

  • 1. 中国科学院重庆绿色智能技术研究院, 自动推理与认知重庆市重点实验室, 重庆 400714;
  • 2. 中国科学院大学, 北京 100049
    基金项目: 国家自然科学基金(批准号:11301524)、重庆市科技攻关项目(批准号:cstc2012ggB40004)和中国科学院西部之光联合学者资助的课题.

摘要: 在计算关联维数过程中, 为了减少人为因素识别无标度区间带来的误差, 提出一种基于模拟退火遗传模糊C均值聚类识别无标度区间的新方法. 该方法根据无标度区间对应曲线的二阶导数在零附近波动的变化特征, 利用分类算法进行识别. 首先对双对数关联积分的离散数据进行二阶差分; 然后利用模拟退火遗传模糊C均值聚类方法对该数据进行分类, 选出在零附近波动的数据; 再剔除粗大误差保留有效数据; 最后进行统计分析识别出线性度最好的作为无标度区间. 应用新方法对两个著名的混沌系统Lorenz 和Henon 进行了仿真, 计算结果与理论值非常符合. 实验表明, 所提出的新方法与主观识别、K-means和2-means方法比较, 可以有效自动识别无标度区间, 减少误差, 计算结果更加精确.

English Abstract

参考文献 (29)

目录

    /

    返回文章
    返回