|
|
Multivariate time series classification based on μσ-DWC feature and tree-structured M-SVM |
TAN Hailong1, LIU Kangling1, JIN Xin1, SHI Xiangrong2, LIANG Jun1 |
1. Department of Control Science and Engineering, Zhejiang University, Hangzhou 310027, China; 2. Department of Information Management, Zhejiang University of Finance and Economics, Hangzhou 310018, China |
|
|
Abstract Aiming at the realization of multivariate time series classification, a feature extraction method based on dimension statistics-wavelet coefficients (μσ-DWC) and a new classification model based on multi-class support vector machine (M-SVM) with tree structure were proposed. The classification algorithm was realized as follows. Firstly, map original multivariate time series to feature space by the proposed feature extraction method. The compressed representation of original time series, namely eigenvector, was obtained. Secondly, tree-structured M-SVM model was trained and constructed after getting the eigenvector representation of training set. Finally, the eigenvector of an unknown time series was extracted and put into the trained tree-structured M-SVM model. Thus, the classification process was completed and the class label of the unknown time series was obtained. The experimental results showed that the proposed algorithm provided higher classification accuracy and faster prediction speed than traditional classification methods, and gave relatively ideal training speed at the same time.
|
Published: 01 June 2015
|
|
基于μσ-DWC特征和树结构M-SVM的多维时间序列分类
为了实现多维时间序列的分类,提出基于统计量-小波系数(μσ-DWC)的序列特征提取方法和新型树结构多分类支持向量机M-SVM模型.分类算法的实现过程如下:利用该特征提取方法将原始多维时间序列映射到特征空间,获得原始序列的压缩表示,即特征向量;得到训练集的特征向量表示之后,训练和构建树结构M-SVM模型;提取未知序列的特征向量并输入已训练完成的树结构M-SVM模型,得到未知序列的类标号,完成分类.实验结果表明:该算法比传统的分类方法具有更高的分类准确率和预测速度,同时可以保证较理想的训练速度.
|
|
[1] AGRAWAL R, FALOUTSOS C, SWAMI A. Efficient similarity search in sequence databases[M]. Springer Berlin Heidelberg: Springer, 1993: 69-84.
[2] CHAN K P, FU A W C. Efficient time series matching by wavelets [C]∥ Proceedings of the International Conference on Data Engineering, Washington: IEEE, 1999: 126-133.
[3] KORN F, JAGADISH H V, FALOUTSOS C. Efficiently supporting ad hoc queries in large datasets of time sequences[J]. ACM SIGMOD Record, 1997, 26(2): 289-300.
[4] KEOGH E J, SMYTH P. A probabilistic approach to fast pattern matching in time series databases [C]∥Proceedings of 3rd International Conference on Knowledge Discovery and Data Mining, Newport Beach: AAAI, 1997: 24-30.
[5] MISHRA S, BHENDE C N, PANIGRAHI K B. Detection and classification of power quality disturbances using Stransform and probabilistic neural network[J]. IEEE Transactions on Power Delivery, 2008, 23(1): 280-287.
[6] MADZAROV G, GJORGJEVIKJ D, CHORBEV I. A multiclass SVM classifier utilizing binary decision tree[J]. Informatica, 2009: 233-241.
[7] 刁智华,赵春江,郭新宇,等.一种新的基于平衡决策树的SVM多分类算法[J].控制与决策,2011,26(1): 149-152.
DIAO Zhihua, ZHAO Chunjiang, GUO Xinyu, et al. A new SVM multiclass classification algorithm based on
balance decision tree [J].Control and Decision, 2011,26 (1): 149-152.
[8] ERI??塁TI H, UAR A, DEMIR Y. Waveletbased feature extraction and selection for classification of power system disturbances using support vector machines [J].Electric Power Systems Research, 2010, 80(7): 743-752.
[9] CHAN K P, FU A W C. Efficient time series matching by wavelets [C]∥IEEE International Conference on Data Engineering, \[S.l.\]: IEEE Computer Society, 1999: 126-133.
[10] 李航.统计学习方法[M].北京:清华大学出版社. 2012: 96-124.
[11] CHANG C, LIN C. LIBSVM: A library for support vector machines[J]. ACM Transactions on Intelligent Systems and Technology, 2001, 2(3): 389396.
[12] 葛哲学.神经网络理论与Matlab R2007实现[M].北京:电子工业出版社.2007: 126-127.
[13] MOHAMMED W K. Highquality recordings of Australian sign language signs[EB/OL]. [20020226]. http:∥archive.ics.uci.edu/ml/machinelearningdatabases/auslan2mld/auslan.html.
[14] MOULDI B. Spoken Arabic digit[EB/OL].[20100913]. http:∥archive.ics.uci.edu/ml/machinelearningdatabases/00195/. |
|
Viewed |
|
|
|
Full text
|
|
|
|
|
Abstract
|
|
|
|
|
Cited |
|
|
|
|
|
Shared |
|
|
|
|
|
Discussed |
|
|
|
|