An Intrusion Detection System (IDS) using multiple detectors and based on Negative Selection Algorithm (NSA)

According to the increasing of network usage in the world, security could be mentioned as an important and valuable issue in communication world. In order to achieve higher security, a lot of researches have been done in the last two decades. Network-based intrusion detection systems (NIDSs) are considered as a prominent challenge in network security subject. There are several algorithms and methods have been introduced up to now. Artificial Immune System (AIS) is one of the most used one. Totally AIS inspired by human natural immune system. The goal of this paper is to increase U2R and R2L classification accuracy which usually have lower accuracy than the others that is DoS and PROBE. In our proposed algorithm, a new set named abnormal set added to the training phase and also a new fitness function is proposed in our Genetic Algorithm (GA). All experiments are performed using KDDCUP99 dataset and the experimental results showed the higher accuracy in both anomaly detection and attack classification of the proposed algorithm compared to recent provided IDS algorithms

کلیدواژه ها:

Network ، Intrusion Detection System (IDS) ، Artificial Immune System (AIS) ، Detectors ، U2R ، R2L