Anomaly-Based Intrusion Detection System Using Support Vector Machine

Abstract

In recent years, there has been an increase in digitization of records, from small databases like student details in a school and product inventory in a shop to large databases like Social Security Number in a country. This digitization even though takes lesser space than its analogy counterparts is susceptible to attacks even from remote locations. Nowadays, there has been a substantial increase in the number of cases of anomalous activities in the network which threatens network safety. So, it is important to not only store the data but also collect the session details so as to distinguish between a normal session and an abnormal session. In this paper, we propose an effective anomaly detection system for cloud computing. The support vector machine is used for profile training and intrusion detection. Experimental results show that IDS with an optimized NSL-KDD dataset using the best feature set algorithm based on Information Gain Ratio increases the accuracy of 96.24% and minimizes the false alarm rate. The machine learning-based approach such as support vector machine has significant potential benefits for the evolution of IDS programs for challenging complex environments such as cloud computing.