吉林大学学报(工学版) ›› 2013, Vol. 43 ›› Issue (05): 1415-1426.doi: 10.7964/jdxbgxb201305042

• 论文 • 上一篇    下一篇

基于半监督boosting表面肌电信号多类模式识别

李阳1, 田彦涛2,3, 陈万忠2   

  1. 1. 北京石油化工学院 信息工程学院, 北京 102617;
    2. 吉林大学 通信工程学院, 长春 130022;
    3. 吉林大学 工程仿生教育部重点实验室, 长春 130022
  • 收稿日期:2012-05-14 出版日期:2013-09-01 发布日期:2013-09-01
  • 通讯作者: 田彦涛(1958- ),男,教授,博士生导师.研究方向:复杂系统优化与控制.E-mail:tianyt@jlu.edu.cn E-mail:tianyt@jlu.edu.cn
  • 作者简介:李阳(1984- ),女,讲师,博士.研究方向:生物医学信号处理.E-mail:lyang@bipt.edu.cn
  • 基金资助:

    吉林省科技发展计划项目(20090350);吉林大学"985工程"项目;高等学校博士学科点专项科研基金项目(20100061110029);吉林大学博士研究生交叉学科科研计划项目(2011J009).

sEMG multi-class pattern recognition based on semi-supervised boosting algorithm

LI Yang1, TIAN Yan-tao2,3, CHEN Wan-zhong2   

  1. 1. Information Engineering College, Beijing Institute of Petrochemical Technology, Beijing 102617, China;
    2. College of Communication Engineering, Jilin University, Changchun 130022, China;
    3. Key Laboratory of Bionic Engineering, Ministry of Education, Jilin University, Changchun 130022, China
  • Received:2012-05-14 Online:2013-09-01 Published:2013-09-01

RICH HTML

0   

PDF (PC)

217

摘要：

针对表面肌电信号较为复杂,且获取标注样本代价较大的问题,提出了基于半监督boosting学习的表面肌电信号多类模式识别方法。与目前半监督boosting算法着重解决两类分类问题,将多类问题转化为多个两类问题不同,本文方法通过联合分类置信度及样本间相似度确定每次迭代过程中未标注样本的预测类别,达到利用未标注样本提高多类问题正确识别率的目的,避免了将某一样本划分多类的问题。由实验分析可知,本文算法与现有半监督boosting算法相比,正确识别率更高,对于不同标注样本数及不同基分类器具有较好的鲁棒性。本文方法降低了人工标注代价,对多类问题具有良好的识别效果。

关键词: 信息处理技术, 肌电信号, 半监督算法, boosting, 多类模式识别

Abstract:

The surface electromyographic signal (sEMG) is often complicated, and it is expensive and time-consuming to obtain labeled samples of sEMG, especially when it has to be done manually by experts. To overcome such problems, a new semi-supervised boosting algorithm is used to classify multi-class problem of sEMG in this paper. This algorithm can use a large number of unlabeled samples together with a small number of labeled samples to build a better learner. Most semi-supervised algorithms have been designed for binary classification. The shortcoming is that when extended to multi-class classification these algorithms are unable to exploit the fact that each sample is only assigned to one class. The advantage of the new semi-supervised boosting algorithm used in this paper is that it exploits both classification confidence and similarities among samples when deciding the pseudo-labels for unlabeled samples. Empirical study with six movements of sEMG show that the new algorithm performs better than the state-of-the-art boosting algorithms for semi-supervised learning. It gives large reduction in the number of human labeled samples, high classification results, which has practical significance in sEMG pattern recognition

Key words: information processing technology, sEMG, semi-supervised algorithm, boosting, multi-class classification

中图分类号: 

  • TN911.7

[1] 刘利, 刘萍萍, 韦佳. 用于带边信息人脸数据的半监督维数约减算法[J]. 吉林大学学报:工学版, 2011, 41(1): 189-193. Liu Li, Liu Ping-ping, Wei Jia. Semi-supervised dimensionality reduction algorithm applying in face data with side information[J]. Journal of Jilin Universit (Engineering and Technology Edition), 2011, 41(1): 189-193.

[2] 董元方, 李雄飞, 李军, 等. XML文档分类的IL-Adaboost算法[J]. 吉林大学学报:工学版, 2011, 41(4): 1054-1058. Dong Yuan-fang, Li Xiong-fei, Li Jun, et al. IL-adaboost algorithm for XML document classification[J]. Journal of Jilin University(Engineering and Technology Edition), 2011, 41(4): 1054-1058.

[3] Bennett K P, Demiriz A. Semi-supervised support vector machine[C]//NIPS, Denver, USA: IEEE, 1999:368-374.

[4] Chapelle O, Zien A. Semi-supervised classification by low fensity deperation[C]//10th Int Workshop on AI and Stat, USA: IEEE, 2005:57-64.

[5] Blum A, Chawla S. Learning from labeled and unlabeled data using graph mincuts[C]//ICML CA, USA: IEEE, 2001:19-26.

[6] Zhu X, Ghahramani Z, Lafferty J. Semi-supervised learning using Gaussian fields and Harmonic functions[C]//ICML, Washington D C, USA: IEEE, 2003: 58-65.

[7] Zhou D, Bousquet O, Lal T, et al. Scholkopf B learning with local and global vonsistency[C]//COLT Cambridge, MA: MIT Press, 2004:321-328.

[8] Belkin M, Matveeva I, Niyogi P. Regularization and semisupervised learning on large graphs[C]//COLT,2004: 624-638.

[9] Bennett K P, Demiriz A, Maclin R. Exploiting unlabeled fata in ensemble methods[C]//KDD,Edmonton, Canada. 2002:289-296.

[10] Chen K, Wang S. Regularized boost for semi-supervised learning[C]//NIPS,2008:281-288.

[11] Jin R, Zhang J. Multi-class learning by smoothed boosting[J]. Mach Learn,2007, 67(3): 207-227.

[12] Scholkopf B, Smola A J. Learning with kernels: support vector machines, regularization, optimization, and Beyond[P]. MIT Press, Cambridge, MA,2002.

[13] Zadrozny B, Elkan C. Transforming classifier scores into accurate multiclass probability estimates[C]//KDD,New York, USA: ACM Press, 2002:694-699.

[14] Dalche Buc F, Grandvalet Y, Ambroise C. Semi-supervised margin boost//NIPS, Cambridge, MA: MIT Press, 2002:553-560.

[15] Higham N J. Matrix nearness problems and applications: Applications of matrix theory[P].Oxford University Press,1989: 1-27.

[16] 李阳, 田彦涛, 陈万忠. 基于FFT盲辨识的肌电信号建模及模式识别[J]. 自动化学报, 2011, 38(1):128-134. Li Yang, Tian Yan-tao, Chen Wan-zhong. Modeling and classifying of sEMG based on FFT blind identification[J]. Acta Automatica Sinica, 2011, 38(1):128-134.

[17] Leo G. Gradient boosting trees for auto insurance loss cost modeling and prediction[J].Expert Systems with Applications, 2012, 39: 3659-3667.
[1] 苏寒松,代志涛,刘高华,张倩芳. 结合吸收Markov链和流行排序的显著性区域检测[J]. 吉林大学学报(工学版), 2018, 48(6): 1887-1894.
[2] 徐岩,孙美双. 基于卷积神经网络的水下图像增强方法[J]. 吉林大学学报(工学版), 2018, 48(6): 1895-1903.
[3] 黄勇,杨德运,乔赛,慕振国. 高分辨合成孔径雷达图像的耦合传统恒虚警目标检测[J]. 吉林大学学报(工学版), 2018, 48(6): 1904-1909.
[4] 李居朋,张祖成,李墨羽,缪德芳. 基于Kalman滤波的电容屏触控轨迹平滑算法[J]. 吉林大学学报(工学版), 2018, 48(6): 1910-1916.
[5] 应欢,刘松华,唐博文,韩丽芳,周亮. 基于自适应释放策略的低开销确定性重放方法[J]. 吉林大学学报(工学版), 2018, 48(6): 1917-1924.
[6] 陆智俊,钟超,吴敬玉. 星载合成孔径雷达图像小特征的准确分割方法[J]. 吉林大学学报(工学版), 2018, 48(6): 1925-1930.
[7] 刘仲民,王阳,李战明,胡文瑾. 基于简单线性迭代聚类和快速最近邻区域合并的图像分割算法[J]. 吉林大学学报(工学版), 2018, 48(6): 1931-1937.
[8] 单泽彪,刘小松,史红伟,王春阳,石要武. 动态压缩感知波达方向跟踪算法[J]. 吉林大学学报(工学版), 2018, 48(6): 1938-1944.
[9] 姚海洋, 王海燕, 张之琛, 申晓红. 双Duffing振子逆向联合信号检测模型[J]. 吉林大学学报(工学版), 2018, 48(4): 1282-1290.
[10] 全薇, 郝晓明, 孙雅东, 柏葆华, 王禹亭. 基于实际眼结构的个性化投影式头盔物镜研制[J]. 吉林大学学报(工学版), 2018, 48(4): 1291-1297.
[11] 陈绵书, 苏越, 桑爱军, 李培鹏. 基于空间矢量模型的图像分类方法[J]. 吉林大学学报(工学版), 2018, 48(3): 943-951.
[12] 陈涛, 崔岳寒, 郭立民. 适用于单快拍的多重信号分类改进算法[J]. 吉林大学学报(工学版), 2018, 48(3): 952-956.
[13] 孟广伟, 李荣佳, 王欣, 周立明, 顾帅. 压电双材料界面裂纹的强度因子分析[J]. 吉林大学学报(工学版), 2018, 48(2): 500-506.
[14] 林金花, 王延杰, 孙宏海. 改进的自适应特征细分方法及其对Catmull-Clark曲面的实时绘制[J]. 吉林大学学报(工学版), 2018, 48(2): 625-632.
[15] 王柯, 刘富, 康冰, 霍彤彤, 周求湛. 基于沙蝎定位猎物的仿生震源定位方法[J]. 吉林大学学报(工学版), 2018, 48(2): 633-639.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
版权所有 © 《吉林大学学报(工学版)》编辑部
地址：长春市人民大街5988号　电话：+86-431-85095297　传真：+86-431-85094074　E-mail: xbgxb@jlu.edu.cn
本系统由北京玛格泰克科技发展有限公司设计开发