Applying Artificial Intelligence Techniques on Cyber Security Datasets: Detecting Cyber Attacks.

Abstract

The rapid expansion of government and corporate services to the online sphere has spurred a notable surge in internet usage among individuals. However, this increased connectivity also amplifies the risks posed by cyber threats, as hackers exploit external networking avenues and corporate networks for personal activities. Consequently, proactive measures must be taken to mitigate potential financial losses and resource drain from cyber attacks. To this end, numerous machine-learning techniques have been developed for cybercrime detection and threat mitigation. This study evaluates several prominent machine learning methods to identify and address significant cyber threats. The research scrutinizes the effectiveness of five techniques: Random Forest, Decision Tree, Convolutional Neural Network (CNN), K-Nearest Neighbors (KNN), and Naive Bayes. Among these, Random Forest demonstrates superior performance with an accuracy rate of 99.69%, outperforming ensemble models such as Decision Tree, CNN, KNN, and Naive Bayes.