ACO-based Intrusion Detection Method in Computer Networks using Fuzzy Association Rules
S. Miryahyaie, Hossein Ebrahimpour-Komleh, Ali Mohammad Nickfarjam · 2021
With the increased use of computer networks, detection of malicious activities and intrusion into systems and computer networks is a basic challenge in this area. The proposed method of this study for intrusion detection using fuzzy association rules and ant colony optimization includes nine basic steps. Ant colony is executed on a graph that each node in the graph provides seven parameters for the proposed method. The initial fuzzy rules are developed according to the values in the most populous city in the graph related to the ant colony by Apriori algorithm. The primary rules are optimized using a procedure called conventional weighted aggregation so as to maximize the accuracy of intrusion detection in computer networks. The optimized rules' efficiency is evaluated using decision tree on NSL-KDD data. Evaluations indicate that the proposed method is of high accuracy in intrusion detection and reduction of error risk compared to competitor methods.