A novel approach to intrusion detection using SVM ensemble with feature augmentation

Network security has been a very important problem. Intrusion detection systems have been widely used to protect network security. Various machine learning techniques have been applied to improve the performance of intrusion detection systems, among which ensemble learning has received a growing interest and is considered as an effective method. Besides, the quality of training data is also an essential determinant that can greatly enhance the detection capability. Knowing that the marginal density ratios are the most powerful univariate classifiers. In this paper, we propose an effective intrusion detection framework based on SVM ensemble with feature augmentation. Specifically, the logarithm marginal density ratios transformation is implemented on the original features with the goal of obtaining new and better-quality transformed training data; SVM ensemble was then used to build the intrusion detection model. Experiment results show that our proposed method can achieve a good and robust performance, which possesses huge competitive advantages when compared to other existing methods in terms of accuracy, detection rate, false alarm rate and training speed.