Distributed-Interval Type-2 Fuzzy Set Based Recognition Algorithm for IDS

In real world, administrators recognize system intrusions through a procedure that can be best explained by the granular computing concepts. This is due to the fact that discrimination between normal and abnormal behavior is not sharply defined. Moreover, the granules in which the incoming requests may be granulated are neither disjoint nor well-defined. This signifies the existence of degrees of uncertainty in the process of intrusion detection. Toward the aim of manipulating and minimizing the effects of uncertainties in the system, we have proposed an algorithm that, based on distributed-interval type-2 fuzzy sets, analyses anomalous behavior trends of system parameters. Based on the analysis results, the system's incoming requests would be treated accordingly.