Research on Attribute Reduction Algorithms in the Intrusion Detection System
Yongzhong Li · Computer and Digital Engineering · 2012
In order to obtain an optimal attribute reduction with larger attribute dependency and fewer attributes number in the large intrusion detection dataset,a new attribute reduction algorithm is proposed in this paper,which combined with rough sets and quantum particle swarm optimization(QPSO).This algorithm solves the disadvantage of traditional way of requiring a lot of labeled data,through define a appropriate fitness function using rough set theorem can make it be implemented in a small number of labeled data.The simulation results on KDDCUP99 dataset show that this algorithm can not only obtain an optimal attribute reduction with fewer attributes number,and the detection accuracy is better than other algorithm.