Generating artificial attack data for intrusion detection using machine learning

Intrusion detection based upon machine learning is currently attracting considerable interests from the research community. One of the appealing properties of machine learning based intrusion detection systems is their ability to detect new and unknown attacks. In order to apply machine learning to intrusion detection, a large number of both attack and normal data samples need to be collected. While, it is often easier to sample benign data based on the normal behaviors of networks, intrusive data is much more scarce, therefore more difficult to collect. In this paper, we propose a novel solution to this problem by generating artificial attack data for intrusion detection based on machine learning techniques. Various machine learning techniques are used to evaluate the effectiveness of the generated data and the results show that the data set of synthetic attack data combining with normal one can help machine learning methods to achieve good performance on intrusion detection problem. Copyright 2014 ACM.