Light Weight Intrusion Detection System with Wrapper Approach and Optimized Feature Selection
Features based Intrusion Detection Systems (IDS), mostly used for Denial of Service (DOS) attacks, have low response in terms of intrusion detection because of missing Local Area Network Denial (LAND) and duration features. Hence, precise security of a system is not assured without considering LAND and duration features. In order to minimize DOS attacks and to make the system more secured, it warrants additional features. All the features are having their values that indicate the presence or absence ofan intrusion. An existing genetic algorithm has considered 16 features for intrusion detection but, still some DOS &Remote to Local (R2L) attacks are not covered in it. These attacks are depends on duration &LAND features of dataset. These two features are focused and extracted using genetic algorithm so that detection response of IDS’s is improved.