Application of Attribute Reduction Based on QPSO in NIDS

The SVM is one of the most successful classification algorithms in network intrusion detection (NIDS) area, but its long training time limits its use.This paper presents a method for enhancing the training time of SVM using attribute reduction optimized by quantum-behaved particle swarm optimization (QPSO), specifically when dealing with large training data sets in NIDS.The reduction algorithm based on attribute reduction optimized by QPSO is used to eliminate the redundant features of sample data set, with the attributes of the raw data are reduced, the SVM training time are reduced.The NIDS based on attribute reduction optimized by QPSO and SVM has better performance.Experimental results show that this method is efficient.