AN INTRUSION DETECTION MODEL BASED ON FUZZY MEMBERSHIP FUNCTION USING GNP

Mahadevappa Immannavar, Prasad Pujar, Manjunath Suryavanshi

Abstract: As the Internet facilities increasing over the world, threats, attacks or intrusions over the Internet are also increasing. Therefore, an intrusion detection model is required to detect intrusion that going to threaten CIA of internet resources. A GNP based fuzzy membership function is much more suitable for identifying such kind of intrusions. A GNP which is a combination of GA and GP applied to extract association rules. A combined GNP-fuzzy membership method would help us to extract important association rules from DARPA 98/99 dataset rather than all rules from DARPA 98/99 dataset. Then the extracted association rules would be updated using genetic operations and also stored into rule pool. In classification, association rules will be classified as normal or intrusion based on calculated match degree. The classified association rules will be stored separately in two different rule pools. Normal rules in normal rule pool and intrusion rules in intrusion rule pool. For the new data match degree will be calculated based on available normal rules and intrusion rules. Then this calculated match degree will help us to identify whether the new data normal or intrusion.

Keywords: Fuzzy membership function, Genetic network programming, Genetic algorithm, DARPA 98/99 dataset and Intrusion detection.

DOI: https://doi.org/10.15623/ijret.2015.0408005