Information fusion for intrusion detection

Intrusion detection is to monitor and capture intrusions into computer and network systems that attempt to compromise the security of computer and network systems. Different intrusion detection techniques exist to evaluate the likelihood of observed activities as a part of an intrusion. When applied to the same observed activities of computer and network systems, different intrusion detection techniques yield different evaluation results. An information fusion technique is required to fuse different results of various intrusion detection techniques for producing a composite value of intrusion likelihood. This paper examines three information fusion techniques based on artificial neural network, linear regression, and logistic regression. These information fusion techniques are compared with respect to their performance.