计算机科学 ›› 2017, Vol. 44 ›› Issue (6): 189-198.doi: 10.11896/j.issn.1002-137X.2017.06.032
孙强,魏伟,侯培鑫,岳继光
SUN Qiang, WEI Wei, HOU Pei-xin and YUE Ji-guang
摘要: 异常检测是系统运行维护的重要工作。在系统运行过程中可获得大量正常的运行数据,但异常数据的获取成本较高,因此可引入单分类器的思想来处理异常检测问题。测量不确定性、环境噪声、存储设备等导致监测数据可能存在不确定性。利用区间数描述不确定的监测数据,提出区间数样本的核可能性1-均值单簇聚类-单分类器异常检测算法。分别考虑聚类中心位于输入空间与特征空间两种情况,并考虑区间数样本具有的区间宽度不均衡性,提出区间细分检测策略。结合人工数据集与UCI数据集给出的算例验证了所提算法的有效性,其与现有SVM-OCC相比具有更高性能。
[1] MOORE R E.Interval arithmetic and automatic error analysis in digital computing[D].Palo Alto:Stanford University,1962. [2] FILIPPONE M,MASULLI F,ROVETTA S.Applying the possibilistic c-means algorithm in kernel induced spaces[J].IEEE Transcations on Fuzzy Systems,2010,8(3):572-584. [3] REN S J,LV J H.Genetic algorithm based kernel function FCM clustering algorithm for interval numbers[J].Journal of System Engineering,2008,3(5):611-616.(in Chinese) 任世锦,吕俊怀.基于遗传算法的区间数核模糊聚类算法[J].系统工程学报,2008,3(5):611-616. [4] PIMENTEL B,COSTA A,SOUZA R.Kernel-based fuzzy clustering of interval data[C]∥Proceedings of 2011 IEEE International Conference on Fuzzy Systems.Taipei,2011:497-501. [5] PIMENTEL B,COSTA A,SOUZA R.Input space versus feature space in kernel-based interval fuzzy C-Means clustering[C]∥Proceedings of 2015 International Joint Conference on Neural Networks.2015:1-7. [6] VAPNIK V N.The Nature of Statistical Learning Theory[M].London:Springer,2000. [7] TAX D M J,DUNI R P W.Support vector domain description[J].Pattern Recognition Letters,1999,0(11):1191-1199. [8] SCHOELKOPF B,SMOLA A J.Learning with kernels:support vector machines,regularization,optimization,and beyond[M].Cambridge,Massachusetts:The MIT Press,2002. [9] CAMPBELL C,BENNETT K P.A linear programming ap-proach to novelty detection[C]∥Proc of the Conference on Neural Information Processing Systems:Natural and Synthetic.Vancouver,Canada,2001:395-401. [10] UTKIN L V,CHEKH A I.A new robust model of one-classclassification by interval-valued training data using the triangular kernel[J].Neural Networks,2015,9:99-110. [11] CARVALHO F,SOUZA R,BEZERRA L.A dynamical clustering method for symbolic interval data based on a single adaptive Euclidean distance[C]∥Proc of the Ninth Brazilian Symposium on Neural Networks (SBRN’06).2006. [12] KRISHNAPURAM R,KELLER J M.A possibilistic approach to clustering[J].IEEE Transcations on Fuzzy Systems,1993,1(2):98-110. [13] ANDERSON D T,BEZDEK J C,POPESCU M,et al.Comparing fuzzy,probabilistic,and possibilistic partitions[J].IEEE Transcations on Fuzzy Systems,2010,8(5):906-918. [14] CHEN B.Research on Outlier Detection Method and Its KeyTechniques[D].Nanjing:Nanjing University of Aeronautics and Astronautics,2013.(in Chinese) 陈斌.异常检测方法及其关键技术研究[D].南京:南京航空航天大学,2013. [15] HEIJDEN F,DUIN R,RIDDER D,et al.Classification,parameter estimation and state estimation-an engineering approach using Matlab[M].Wiley,2004. [16] LICHMAN M.UCI Machine Learning Repository[DB/OL].http://archive.ics.uci.edu/ml. |
No related articles found! |
|