Multi-Objective Fish Swarm Optimization with Fuzzy Association Rule for Botnet Detection System
R. Kiruthika, K. Selvam · 2022 3rd International Conference on Smart Electronics and Communication (ICOSEC) · 2022
Botnets have typically been considered as a major security risk for networks. As the distinctions between regular and botnet traffic grow, botnet detection approaches based on traffic analysis frequently produce substantial false positive rates. BOTNET is a term that encompasses the internet’s high-level complexity. Since the internet is the most important medium of communication, bot generate internet traffic. Botnet covers a wide range of internet problems, including spam, malware, click fraud, and phishing. Algorithms can withstand a variety of DDoS attacks. A variety of machine learning techniques can detect and mitigate these types of threats using optimization methodologies. This paper has designed and implemented a novel method called “Multi-Objective Fish Swarm optimization with Fuzzy Association Rule for Botnet Detection System”. A Cloud wick interface is used to detect the live bots in Virtual Private Server. Real-time server data such as KDD-Cup and NSL-KDD datasets were used. These datasets are a collection of recorded network data used to support the construction of a new intruder detection system. RSD data is the real time network traffic data fetched through the built-in firewall of a dedicated private server. This implementation has been done by using 11 attributes extracted from the real time network. Out of the 11, 7 attributes are discrete and 4 attributes are continuous. The dataset is monitored every 6 minutes to obtain the results. The accuracy percentage for the Multi-Objective Fish Swarm optimization with Fuzzy Association Rule for Botnet Detection System is 99.4%.