Intrusion Detection Using Ensemble of SVM Classifiers

Abstract

The Current researches show that different classifiers provide different results about the patterns to be classified. These different results combined together yields better performance than individual classifiers. An ideal classifier, which is popularly known as the ensemble approach, is combined to take the final decision instead rely on a single classifier for decision on our intrusion detection system. Weight voting rule, unlike majority voting rule, is a highlight of our ensemble performance. The remarkable highlight is choosing the optimal weights strategy. In the performance of our intrusion detection system, the weight values are based on the accuracy of a given data class actually classified by each classifier respectively. In fact, our experiments show that Intrusion Detection performances can be improved by combining an ensemble of SVM classifiers.