Effective Anomaly based Intrusion Detection using Rough Set Theory and Support Vector Machine

Abstract

Intrusion detection system is used to discover illegitimate and unnecessary behavior at accessing or manipulating computer systems. Subsequently, these behaviors are checked as an attack or normal behavior. Intrusion detection systems aim to identify attacks with a high detection rate and a low false positive. Most of the earlier IDs make use of all the features in the packet to analyze and look for well-known intrusive models. Some of these features are unrelated and superfluous. The disadvantage of these methods is degrading the performance of IDs. The proposed Rough Set Support Vector Machine (RSSVM) approach is extensively decreases the computer resources like memory and CPU utilization which are required to identify an attack. The approach uses rough set to find out feature reducts sets. Then reduct sets are sent to SVM to train and test data. The results showed that the proposed approach gives better and robust representation of data.