Cyber Attack Classifications Using Supervised Learning Technique

• Published in: 2024 2nd International Conference on Artificial Intelligence and Machine Learning Applications Theme: Healthcare and Internet of Things (AIMLA) pp. 1 - 7
• Main Authors: Infantia H, Niroshini, M, Revathi, WU, Rechu, Sneha
• Format: Conference Proceeding
• Language: English
• Published: IEEE 15.03.2024
• Subjects: Analytical models; Classification; Cleaning; Cyber Attacks; Data models; Datasets; Feature Extraction; Machine learning; Telecommunication traffic; Threat assessment; Training
• DOI: 10.1109/AIMLA59606.2024.10531367

The unauthorized and malevolent action targeted at taking advantage of weakness surrounded by electronic systems, networks or appliances characterize cyber attacks. These attacks are designed to disrupt, access unauthorized information, cause damage, or gain control over sensitive data. They encompass a wide array of tactics, including Malware, Zero-Day Exploits, Denial-of-Service (DoS), Cross-Site Scripting(XSS), SQL injections, and ransomware, among others. Their impact can be devastating, leading to financial loss, compromised data integrity, and damage to an individual's or organization's reputation. As technology evolves, cyber attacks continue to advance in sophistication and frequency, posing a significant threat to cyber security. Preventative measures and robust security protocols are vital to combat these threats. The proactive identification and mitigation of cyber attacks are crucial in safeguarding digital assets, requiring continuous vigilance, up-to-date security mechanisms, and an understanding of emerging threats. Addressing these challenges is essential for the protection of digital infrastructure and the preservation of individuals' and organizations' sensitive information and assets.The project aims to create a system using supervised machine learning techniques to categorize diverse cyber attacks. It utilizes a well-structured datasets encompassing various attack types like malware, phishing, and DDoS attacks. This datasets includes network traffic data and behavioural attributes for feature extraction, enabling the training of a strong classification model. The model will be trained using labelled historical instances of cyber attacks, enabling it to identify complex patterns and distinctions among different attack types. Regular model updates and retraining with new attack data are proposed to maintain the model's relevance in the constantly evolving threat landscape. The anticipated outcome is an effective system capable of accurately identifying and categorizing cyber threats.