基于深度学习的时间序列数据异常检测方法
|
摘要
针对类间分布不平衡的时间序列数据的异常检测问题,提出了一种基于深度卷积神经网络的检测方法.首先采用抽样法对不平衡时间序列数据进行预处理;其次,将处理后的时间序列数据转换为尺度一致、时长一致的片段;最后将数据送入具有4层隐藏层结构的卷积神经网络模型中进行异常检测.实验结果表明,所提方法弥补了现存的检测技术由于忽略数据分布的偏斜性而造成的少数类检测精度低的缺点,并通过与现有的时间序列分类方法的比较,验证了所提方法的高效性.
With regard to the anomaly detection problem of time-series data with a skewed between-class distribution,we propose a detection method based on deep convolutional neural network. First,we employ the sampling method to preprocess the unbalanced time-series data. Second,the original time-series data are converted into a series of continuous segments with a uniform scale and consistent duration. Finally,we feed the data into a convolutional neural network model with four hidden layers for anomaly detection. The experimental results show that the proposed method covers the shortage of existing detection technologies that ignore the skewness of data distributions and results in a low-detection precision. Compared with the existing time-series classification methods,the proposed method provided a satisfactory performance.
With regard to the anomaly detection problem of time-series data with a skewed between-class distribution,we propose a detection method based on deep convolutional neural network. First,we employ the sampling method to preprocess the unbalanced time-series data. Second,the original time-series data are converted into a series of continuous segments with a uniform scale and consistent duration. Finally,we feed the data into a convolutional neural network model with four hidden layers for anomaly detection. The experimental results show that the proposed method covers the shortage of existing detection technologies that ignore the skewness of data distributions and results in a low-detection precision. Compared with the existing time-series classification methods,the proposed method provided a satisfactory performance.
引文
[1]林智勇,郝志峰,杨晓伟.不平衡数据分类的研究现状[J].计算机应用研究,2008,25(2):332-336.Lin Z Y,Hao Z F,Yang X W. Research status of unbalanced data classification[J]. Application Research of Computers,2008,25(2):332-336.
[2]王晓娟,郭躬德.不平衡数据采样方法的对比学习[J].微计算机信息,2011(12):155-157.Wang X J,Guo G D. Comparative study of unbalanced data sampling method[J]. Microcomputer Information,2011(12):155-157.
[3] Li J,Li H,Yu J L. Application of random-SMOTE on imbalanced data mining[C]//Fourth International Conference on Business Intelligenceand Financial Engineering. Piscataway,NJ,USA:IEEE,2011:130-133.
[4] Chawla N V,Bowyer K W,Hall L O,et al. SMOTE:Synthetic minority over-sampling technique[J]. Journal of artificial Intelligence Re-search,2002,16:321-357.
[5]许丹丹,王勇,蔡立军.面向不均衡数据集的ISMOTE算法[J].计算机应用,2011,31(9):2399-2401.Xu D D,Wang Y,Cai L J. ISMOTE algorithm for unbalanced data sets[J]. Journal of Computer Applications,2011,31(9):2399-2401.
[6] Cao H,Li X L,Woon Y K,et al. Integrated oversampling for imbalanced time series classification[J]. IEEE Transactions on Knowledge andData Engineering,2013,25(12):2809-2822.
[7] Ding H,Trajcevski G,Scheuermann P,et al. Querying and mining of time series data[J]. Proceedings of the Vldb Endowment,2008,1(2):1542-1552.
[8] Keogh E,Kasetty S. On the need for time series data mining benchmarks:A survey and empirical demonstration[J]. Data Mining&Knowl-edge Discovery,2003,7(4):349-371.
[9] Yang Y,Deng Q,Shen F,et al. A shapelet learning method for time series classification[C]//International Conference on TOOLS with Artifi-cial Intelligence. Piscataway,NJ,USA:IEEE,2017:423-430.
[10]孙其法,闫秋艳,闫欣鸣.基于多样化top-k shapelets转换的时间序列分类方法[J].计算机应用,2017,37(2):335-340.Sun Q F,Yan Q Y,Yan X M. Classification of time series based on diversified top-k shapelets conversion[J]. Journal of Computer Applica-tions,2017,37(2):335-340.
[11]周飞燕,金林鹏,董军.卷积神经网络研究综述[J].计算机学报,2017,40(6):1229-1251.Zhou F Y,Jin L P,Dong J. An overview of convolutional neural network research[J]. Chinese Journal of Computers,2017,40(6):1299-1251.
[12]Cui Z,Chen W,Chen Y. Multi-Scale convolutional neural networks for time series classification[J/OL]. ar Xiv:1603.06995.[2018-01-20]. http://xueshu.baidu.com/s?wd=paperuri%3A%289f44ff3471953143144883e017f17356%29&filter=sc_long_sign&tn=SE_xueshu-source_2kduw22v&sc_vurl=http%3A%2F%2Farxiv. org%2Fabs%2F1603.06995&ie=utf-8&sc_us=7768111178099589474.
[13]Wang Z,Yan W,Oates T. Time series classification from scratch with deep neural networks:A strong baseline[C]//International Joint con-ference on Neural Networks. Piscataway,NJ,USA:IEEE,2017:1578-1585.
[14]Zhao B,Lu H,Chen S,et al. Convolutional neural networks for time series classification[J]. Journal of Systems Engineering and Electronics,2017,28(1):162-169.
[15]He H,Garcia E A. Learning from imbalanced data[J]. IEEE Transactions on Knowledge and Data Engineering,2009,21(9):1263-1284.
[16]Lekha L,Suchetha M. Real-time non-invasive detection and classification of diabetes using modified convolution neural network[J]. IEEEJournal of Biomedical&Health Informatics,2018,22(5):1630-1636.
[17]Ince T,Kiranyaz S,Eren L,et al. Real-time motor fault detection by 1-D convolutional neural networks[J]. IEEE Transactions on IndustrialElectronics,2016,63(11):7067-7075.
[18]黄文坚,唐源. TensorFlow实战[M]. 1版.北京:电子工业出版社,2017.Hu W J,Tang Y. Tensor Flow practical experience[M]. 1st ed. Beijing:Electronics Industry Press,2017.
[19]Abadi M,Barham P,Chen J,et al. Tensor Flow:A system for large-scale machine learning[C]//Proceedings of the 12th USENIX conferenceon Operating Systems Design and Implementation. Berkeley,CA,USA:USENIX Association,2016:265-283.
[20]李慧.奇异值分解在时间序列分析中的应用[D].北京:北京交通大学,2009.Li H. The application of singular value decomposition in time series analysis[D]. Beijing:Beijing Jiaotong University,2009.
[21]郑恩辉,李平,宋执环.不平衡数据知识挖掘:类分布对支持向量机分类的影响[J].信息与控制,2005,34(6):703-708.Zheng E H,Li P,Song Z H. Unbalanced data knowledge mining:Influence of class distribution on support vector machine classification[J].Information and Control,2005,34(6):703-708.
[2]王晓娟,郭躬德.不平衡数据采样方法的对比学习[J].微计算机信息,2011(12):155-157.Wang X J,Guo G D. Comparative study of unbalanced data sampling method[J]. Microcomputer Information,2011(12):155-157.
[3] Li J,Li H,Yu J L. Application of random-SMOTE on imbalanced data mining[C]//Fourth International Conference on Business Intelligenceand Financial Engineering. Piscataway,NJ,USA:IEEE,2011:130-133.
[4] Chawla N V,Bowyer K W,Hall L O,et al. SMOTE:Synthetic minority over-sampling technique[J]. Journal of artificial Intelligence Re-search,2002,16:321-357.
[5]许丹丹,王勇,蔡立军.面向不均衡数据集的ISMOTE算法[J].计算机应用,2011,31(9):2399-2401.Xu D D,Wang Y,Cai L J. ISMOTE algorithm for unbalanced data sets[J]. Journal of Computer Applications,2011,31(9):2399-2401.
[6] Cao H,Li X L,Woon Y K,et al. Integrated oversampling for imbalanced time series classification[J]. IEEE Transactions on Knowledge andData Engineering,2013,25(12):2809-2822.
[7] Ding H,Trajcevski G,Scheuermann P,et al. Querying and mining of time series data[J]. Proceedings of the Vldb Endowment,2008,1(2):1542-1552.
[8] Keogh E,Kasetty S. On the need for time series data mining benchmarks:A survey and empirical demonstration[J]. Data Mining&Knowl-edge Discovery,2003,7(4):349-371.
[9] Yang Y,Deng Q,Shen F,et al. A shapelet learning method for time series classification[C]//International Conference on TOOLS with Artifi-cial Intelligence. Piscataway,NJ,USA:IEEE,2017:423-430.
[10]孙其法,闫秋艳,闫欣鸣.基于多样化top-k shapelets转换的时间序列分类方法[J].计算机应用,2017,37(2):335-340.Sun Q F,Yan Q Y,Yan X M. Classification of time series based on diversified top-k shapelets conversion[J]. Journal of Computer Applica-tions,2017,37(2):335-340.
[11]周飞燕,金林鹏,董军.卷积神经网络研究综述[J].计算机学报,2017,40(6):1229-1251.Zhou F Y,Jin L P,Dong J. An overview of convolutional neural network research[J]. Chinese Journal of Computers,2017,40(6):1299-1251.
[12]Cui Z,Chen W,Chen Y. Multi-Scale convolutional neural networks for time series classification[J/OL]. ar Xiv:1603.06995.[2018-01-20]. http://xueshu.baidu.com/s?wd=paperuri%3A%289f44ff3471953143144883e017f17356%29&filter=sc_long_sign&tn=SE_xueshu-source_2kduw22v&sc_vurl=http%3A%2F%2Farxiv. org%2Fabs%2F1603.06995&ie=utf-8&sc_us=7768111178099589474.
[13]Wang Z,Yan W,Oates T. Time series classification from scratch with deep neural networks:A strong baseline[C]//International Joint con-ference on Neural Networks. Piscataway,NJ,USA:IEEE,2017:1578-1585.
[14]Zhao B,Lu H,Chen S,et al. Convolutional neural networks for time series classification[J]. Journal of Systems Engineering and Electronics,2017,28(1):162-169.
[15]He H,Garcia E A. Learning from imbalanced data[J]. IEEE Transactions on Knowledge and Data Engineering,2009,21(9):1263-1284.
[16]Lekha L,Suchetha M. Real-time non-invasive detection and classification of diabetes using modified convolution neural network[J]. IEEEJournal of Biomedical&Health Informatics,2018,22(5):1630-1636.
[17]Ince T,Kiranyaz S,Eren L,et al. Real-time motor fault detection by 1-D convolutional neural networks[J]. IEEE Transactions on IndustrialElectronics,2016,63(11):7067-7075.
[18]黄文坚,唐源. TensorFlow实战[M]. 1版.北京:电子工业出版社,2017.Hu W J,Tang Y. Tensor Flow practical experience[M]. 1st ed. Beijing:Electronics Industry Press,2017.
[19]Abadi M,Barham P,Chen J,et al. Tensor Flow:A system for large-scale machine learning[C]//Proceedings of the 12th USENIX conferenceon Operating Systems Design and Implementation. Berkeley,CA,USA:USENIX Association,2016:265-283.
[20]李慧.奇异值分解在时间序列分析中的应用[D].北京:北京交通大学,2009.Li H. The application of singular value decomposition in time series analysis[D]. Beijing:Beijing Jiaotong University,2009.
[21]郑恩辉,李平,宋执环.不平衡数据知识挖掘:类分布对支持向量机分类的影响[J].信息与控制,2005,34(6):703-708.Zheng E H,Li P,Song Z H. Unbalanced data knowledge mining:Influence of class distribution on support vector machine classification[J].Information and Control,2005,34(6):703-708.