An intrusion detection approach inspired by biological memory cell

Nowadays, computers and networks have become the infrastructure of people's daily lives. As the Internet becomes more open to public, most people and organizations prefer to use online services. At the same time, the Internet security problem is becoming more severe. An effective and feasible protection measure for network security is necessary. Intrusion detection system (IDS) is a typical measure used to provide protection for information systems. In order to detect unknown attacks, anomaly detection technique is a commonly used method. Meanwhile, many scholars are studying more effective and easy-to-use anomaly detection methods. In this paper, a new intrusion detection approach inspired by the function of biological memory cell is proposed. The detectors in this approach (which called memory detector) are more effective than common detectors. The generation method of memory detector is detailed. In order to analyze and evaluate the performance of this approach, comparative experiments are implemented with training and testing data from “Knowledge Discovery and Data mining cup 1999” (KDD99) dataset. The results show that our approach provides better performance than ordinary anomaly detection approaches with higher true positive rate and lower false positive rate.