Intrusion Detection System based on Hidden Conditional Random Fields

Intrusion detection is an important way to ensure the security of computers and networks. In this paper, a new intrusion detection system (IDS) is proposed based on Hidden Conditional Random Fields (HCRFs). In order to optimize the performance of HCRFs, we bring forward the Two-stage Feature Selection method, which contains Manual Feature Selection method and Backward Feature Elimination Wrapper (BFEW) method. The BFEW is a feature selection method which is introduced based on wrapper approach. Experimental results on KDD99 dataset show that the proposed IDS not only has a great advantage in detection efficiency but also has a higher accuracy when compared with other well-known methods.