Intrusion detection based on hybrid fuzzy clustering algorithm
Jing Xu · Journal of Jiangsu University of Science and Technology · 2008
Fuzzy clustering algorithm is an unsupervised machine learning method which could detect the unknown and anomaly network intrusions efficiently.But fuzzy clustering is an iterative optimization method and easy to fall into local optimal solution.This paper presented a new hybrid algorithm based on genetic tabu fuzzy clustering algorithm(GTFCM) which combined global search of genetic algorithm with local search of tabu technique.Experimental results show it can avoid to be trapped in local optimal solutions.Tests on the data set of KDDCUP99 show that this algorithm can detect intrusions with higher detection rate and lower false rate.