Research On Ant Clustering Algorithm For Network Intrusion Detection

Resources can be shared fast in computer network, meanwhile, the computer networkitself suffers from all kinds of illegal access and attack frequently. Computer networksecurity is concerned widely. Intrusion detection system is an active safety protectiontechnology. As one of important researches in network security, it has been developedrapidly in recent years.Ant colony algorithm is a simulation evolutionary algorithm based on population. Itseffective bionics process has been widely used in various combinatorial problems, thealgorithm is adaptive and distributed parallel computing and easy to combine with otheralgorithms. The clustering technology is an important area of research content on datamining, the ant colony algorithm combined with cluster analysis techniques was applicationin network intrusion detection technologies, give full play to their advantages, so that theperformance of intrusion detection become more efficient.In this thesis, we studied anomaly intrusion detection technology based on clusteringanalysis. The specific contents of this dissertation are listed as follows:Firstly, we comprehensive and systematic expatiated on current intrusion detectiontechniques and clustering algorithm; analyzed advantages and disadvantaged of existingclustering algorithm which id applied to intrusion detection.Secondly, we describes related technologies about network intrusion detectiontechnology, clustering analysis and ant colony clustering algorithm applied to networkintrusion detection and the process of development and problems, and on this basis, it willintroduced variation factor which was in the genetic algorithms into the ant colonyclustering algorithm, the convergence time can be shortened by variation factor and it willavoid the algorithm into a local optimum.Finally, we use KDD cup99data set which was used in today’s accepted data set inintrusion detection algorithms, the data which was pre-treatment, then use clusteridentification technique which can identify a small number of invasion data, It is provedthat the algorithms we developed can improve the detection rate of the four attack rate andfurther evidence of the improved algorithm has a good clustering results.