A concept lattice-based adaptive intrusion detection algorithm

An adaptive intrusion detection approach based on concept lattices and intrusion detection technology is presented. Firstly, the incremental algorithm is use to construct concept lattices. Secondly, by analyzing the implication relations among the connotation of concepts, detection rules can be extracted and trained repeatedly to construct the intrusion detector. When the false alarm rate or missed rate reaches the set threshold, the concept lattices should be updated to dynamically renew the rule library, and the intrusion detector should be modified to detect new intrusions, in order to detect a variety of known and unknown network attacks effectively. Experiments show that the accuracy of intrusion detection of the proposed method has improved significantly.