Intrusion Detection Based on Information Entropy of Multiple Support Vector Machine

Abstract—Intrusion detection is essentially a classification problem. It is very important to increase the classification accuracy. Support vector machine (SVM) is a kind of very effective tool to solve classification problems, accuracy of SVM intrusion detection Based on information entropy is relatively high, but how to achieve higher accuracy is still a new issue. In this paper, we use multiple SVM intrusion detection algorithm based on information entropy to solve the problem, first, we use information entropy theory to optimize the SVM kernel function, then reduce and reclassify the Characteristic attributes of each Network connection, finally, training each Reclassification set of attributes with two Support vector machine, the experimental results show that the proposed method is an efficient network intrusion detection method. Keywords-Intrusion detection; information entropy; Support vector machine I