Ensemble Methods Classifier Comparison for Anomaly Based Intrusion Detection System on CIDDS-002 Dataset

Abstract

With the rapid development of information technology, the network has been everywhere. This technology has brought a lot of convenience to people, but there are also some security problems. To solve these problems, many methods have been proposed, among which is intrusion detection. A lot of research has been done to find the most effective Intrusion Detection Systems. In term of detecting novel attacks, Anomaly-Based Intrusion Detection Systems has better significance than Misuse-Based Intrusion Detection Systems. The research on the datasets being used for training and testing purposes in the detection model is as important as the model. Better dataset quality can improve intrusion detection model results. This research presents the statistical analysis of labeled flow-based CIDDS-002 dataset using ensemble methods classifier. The analysis is done concerning some prominent evaluation metrics used for evaluating Intrusion Detection Systems including Detection Rate, Accuracy, and False Positive Rate. As a result, the accuracy of the Bagging (Decision Tree) is 99.71% and Bagging (Gaussian Naïve Bayes) is 67.57%.