Intrusion Detection Based on Fuzzy Support Vector Machines
Hongle Du, Teng Shaohua, Zhu Qingfang · 2009
A great deal of noise data in the network connectivity information affect badly to build SVM optimal classification hyperplane and lead to higher classification error rate. In this paper, fuzzy membership function is applied into v-SVM; it acquires different values for each input data that accord to different effects on the classification result. Therefore different input samples points can make different contributions to the learning of the decision surface - the optimal separating hyperplane. Then the model of intrusion detection system based on SVM is presented, and detailedly illustrated the performance of this model. Finally, comparison of detection ability between v-SVM and v-FSVM is given. It is found that v-FSVM effectively reduce the impact of the noise data and improve the accuracy of decision-making.