This paper introduces an original method for

\r\nguaranteed estimation of the accuracy for an ensemble of Lipschitz

\r\nclassifiers. The solution was obtained as a finite closed set of

\r\nalternative hypotheses, which contains an object of classification with

\r\nprobability of not less than the specified value. Thus, the

\r\nclassification is represented by a set of hypothetical classes. In this

\r\ncase, the smaller the cardinality of the discrete set of hypothetical

\r\nclasses is, the higher is the classification accuracy. Experiments have

\r\nshown that if cardinality of the classifiers ensemble is increased then

\r\nthe cardinality of this set of hypothetical classes is reduced. The

\r\nproblem of the guaranteed estimation of the accuracy for an ensemble

\r\nof Lipschitz classifiers is relevant in multichannel classification of

\r\ntarget events in C-OTDR monitoring systems. Results of suggested

\r\napproach practical usage to accuracy control in C-OTDR monitoring

\r\nsystems are present.<\/p>\r\n","references":"[1] A.V. Timofeev, \u201cThe guaranteed estimation of the Lipschitz classifier\r\naccuracy: Confidence set approach.\u201d Journal Korean Stat. Soc. 41(1),\r\n2012, pp.105-114. [2] L. Breiman, \u201cBagging predictors\u201d, Mach. Learn. 24(2), 1996, pp. 123\u2013\r\n140.\r\n[3] P. Rangel, F. Lozano, E. Garc\u00eda, \u201cBoosting of support vector machines\r\nwith application to editing\u201d, Proceedings of the 4nd International\r\nConference of Machine Learning and Applications ICMLA\u201905, Springer\r\n2005.\r\n[4] C.H. You, K.A. Lee, H. Li, \u201cA GMM supervector Kernel with the\r\nBhattacharyya distance for SVM based speaker recognition\u201d, ICASSP,\r\nIEEE International Conference on Acoustics, Speech and Signal\r\nProcessing, 2009, pp. 4221-4224.\r\n[5] A.V. Timofeev, \u201cThe guaranteed detection of the seismoacoustic\r\nemission source in the C- OTDR systems\u201d, ICDSP 2014: International\r\nConference on Digital Signal Processing Conference Proceedings,\r\nBarcelona, Vol.8 N. 10, 2014, pp. 979 \u2013 982.\r\n[6] U. Luxburg, O. Bousquet, \u201cDistance-based classification with Lipschitz\r\nfunctions\u201d, Journal of Machine Learning Research, 5, 2004, pp. 669-\r\n695.\r\n[7] Kailath T. The Divergence and Bhattacharyya Distance Measures in\r\nSignal Selection, IEEE Transactions on Communication Technology 15\r\n(1), 1967, pp. 52-60.\r\n[8] Timofeev A.V., Egorov D.V., Multichannel classification of target\r\nsignals by means of an SVM ensemble in C-OTDR systems for remote\r\nmonitoring of extended objects, MVML-2014 Conference Proceedings\r\nV.1, Prague, 2014.","publisher":"World Academy of Science, Engineering and Technology","index":"Open Science Index 100, 2015"}