학술논문
Hybrid Dragonfly-Cat Swarm Clustering algorithm with Closed Sequential Pattern Mining for Detection of Malicious Transactions
Document Type
Conference
Source
2022 International Conference on Futuristic Technologies (INCOFT) Futuristic Technologies (INCOFT), 2022 International Conference on. :1-8 Nov, 2022
Subject
Language
Abstract
An Intrusion Detection System (IDS) is a tracking system that recognizes suspicious activity and generates alerts when it discovers it.A current intrusion detection system’s major purpose is to identify signature and anomaly-based invasive attacks that abuse database privileges. In this work, we proposed a hybrid meta-heuristic clustering technique employing the Dragonfly and Cat swarm algorithms for building user profiles based on prior behavior. Furthermore, We explored the use of the CloFAST algorithm to mine association rules, which are then used to calculate the proposed dynamic conformance index, thus detecting fraudulent database transactions in a stream of incoming transactions. Experimental results demonstrate that the proposed technique achieves an accuracy of 99.33 % using a synthetically produced dataset in line with the TPC-C benchmark.