×

Parallel proximal support vector machine and its application in intrusion detection. (English) Zbl 1240.68203

Summary: A new training method based on the parallel proximal support vector machine (PSVM) classification algorithm is proposed. An efficient PSVM and the cascade SVM architecture are applied to reduce the time of training by using the equivalence between the \(\varepsilon\)-support vectors and the original dataset. In addition, a new incremental learning method based on PSVM is used to make the update of the classifier easier. Experiments on the KDD CUP 1999 dataset demonstrate that the training time of our method is 20% less than that of other SVM methods under the conditions of ensuring a low false positive rate and a high detection rate; it can update the classifier effectively by learning the characteristics of new dataset incrementally.

MSC:

68T05 Learning and adaptive systems in artificial intelligence
PDFBibTeX XMLCite