Intrusion Detection Ensemble Algorithm based on Bagging and Neighborhood Rough Set
Intrusion detection data often have some characteristics such as nonlinearity, higher dimension, much redundancy and noise, and partial continuous-attribute. This paper presents a new ensemble algorithm to improve intrusion detection precision. Firstly, it generates multiple training subsets in difference by using bootstrap technology. Then using neighborhood rough sets with different radiuses to make attribute reduction in these subsets, obtained the training subsets with greater difference, while Particle Swarm Optimization is used to optimize parameters of support vector machine in order to get base classifiers with greater difference and higher precision. Finally, the above base classifiers were integrdinedd by weighted synthesis method.