Network Intrusion Detection System Using Stacking and Boosting Ensemble Methods

Abstract

The role of an Intrusion Detection System (IDS) is crucial in monitoring the presence of malicious activities or policy violations. Basically, there are basically two types of Intrusion Detection Systems, namely Host-Based and Network Intrusion Detection systems. The former systems are those that are deployed onto any local system, and when it is built onto a network, it is called Network Intrusion Detection system. As the connectivity between computers is increasing, network security is an important aspect to be taken into account. Various techniques and methodologies are carried out in building efficient systems for intrusion detection. [5] The performance of an IDS depends upon various parameters like precision, rate of detection, accuracy, etc. A good intrusion detection system should have a higher accuracy and detection rate, with fewer false alarms. Many research has tried to build an efficient intrusion detection system in different fields. This work focuses on implementing an anomaly based network intrusion detection system using stacking and boosting ensemble methods. Both the approaches are implemented on the same dataset called NSL-Knowledge Discovery Dataset (NSL-KDD), which is a well-known benchmark in the research of Intrusion Detection techniques.