Evolutionary and immune algorithms applied to association rule mining
SEMCCO'12 Proceedings of the Third international conference on Swarm, Evolutionary, and Memetic Computing
| Hi-index | 0.00 |
In this paper, intrusion detection approaches for relational database systems were studied. An immune based intrusion detection algorithm for relational databases was proposed. According to the algorithm, the data to be detected were encoded into binary strings after preprocessing. The philosophy of negative selection in biological immune systems was utilized to generate immune detectors. Intrusion detection was fulfilled by comparing the strings of audit data with immune detectors. Experiments were designed to verify the effectiveness of the proposed algorithm. Based on the same test data, the detection results of proposed algorithm were compared with those of other two detection algorithms: an association rule mining based detection algorithm and a sequential pattern mining based detection algorithm. The results show that the immune based intrusion detection algorithm for relational databases is more effective than the other two algorithms in reducing the false alarm ratio and promoting correctness ratio.