Algorithm of multi-category SVM incremental learning in application of intrusion detection
作者： Meng Fan-xue , Liu Yan-heng , Lu Rong , Wu Jing
DOI:
关键词:
摘要: This paper proposed a new algorithm of multi-category SVM incremental learning by analyzing the distribution characteristics intrusion detection data. Samples used in were selected measuring distance between samples and their class-centers, they are just those which will most possibly be SVs learning. By several binary-class hyper-planes, zones inhomogeneous divided, is realized. Using this algorithm, quantity training reduced while high rate ensured at same time. The test based on KDDCUP99 dataset. testing result proved that time complexity space can both effectively accuracy won't decrease.
参考文章(7)
S. Knerr, L. Personnaz, G. Dreyfus, Single-layer learning revisited: a stepwise procedure for building and training a neural network NATO Neurocomputing. pp. 41- 50 ,(1990) , 10.1007/978-3-642-76153-9_5
Chih-Jen Lin, Chih-Wei Hsu, Chih-Chung Chang, A Practical Guide to Support Vector Classication 臺北市:國立臺灣大學資訊工程學系. ,(2008)
S. Ruping, Incremental learning with support vector machines international conference on data mining. pp. 641- 642 ,(2001) , 10.1109/ICDM.2001.989589
Vladimir N. Vapnik, The Nature of Statistical Learning Theory ,(1995)
Guo Hui, A sort of support vector machine incremental learning algorithm based on clustering Journal of University of Science and Technology Beijing. ,(2007)
Kong Bo, Incremental support vector machine based on center distance ratio Journal of Computer Applications. ,(2006)
Tao Liang, Fast Incremental SVM Learning Algorithm Based on Active Set Iterations Computer Simulation. ,(2006)