Cyber Security Intrusion Detection Using a Deep Learning Method

• Published in: Mehran University research journal of engineering and technology Vol. 44; no. 1; pp. 69 - 74
• Main Authors: Ullah, Basheer, Massan, Shafiq-ur-Rehman, Rehman, M. Abdul, Khan, Rabia Ali
• Format: Journal Article
• Language: English
• Published: Mehran University of Engineering and Technology 01.01.2025
• Subjects: Algorithms; Cyberterrorism; Data security; Innovations; Methods; Neural networks; Security management; Spyware
• ISSN: 0254-7821; 2413-7219
• DOI: 10.22581/muet1982.3170

The World is moving towards information technology dependence, the cornerstone of which is information security. As the number of active connections becomes large so is the need of security increasing day by day. Presently, billions of devices are connected and every hour 0.46 Million new devices are connected to the web. Hence, due to this huge increase, the number of interconnections and the use of diverse protocols increases. Information and cyber security is a challenge worldwide and a big issue in business. One of the major aspects of information security is intrusion detection. It is important for cyber protection due an increasing number of cyber-attacks. Present methods to detect, predict and prevent malware still fall short of the desired level. The new techniques of deep learning are poised to succeed for detecting intrusion by employing different algorithms of detection and prevention. This paper proposes a deep neural network (DNN) for intrusion detection by the use of Kaggle NLS-KDD dataset with the highest attained accuracy of 92%. This detection method may prove to be very useful for ensuring cyber security of computers hence preventing data and economic loss.