Advancements in Cybersecurity: Evaluating Machine Learning Approaches for Detecting Cyber Attacks

These days, technological advancements are the main cause of the exponential growth in cyberattacks. Cyber-attacks may be detected using AI, ML, and DL approaches that use massive volumes of data. These learning approaches are used to identify a broad range of cyber-attacks by analysing the web traffic or network traffic to identify potential threats. This study explores an effectiveness of various ML models, including ANN, KNN, Multilayer Perceptrons (MLP), and LR, in detecting cyber-attacks using the UNSW-NB15 dataset. A result reveal that an ANN model significantly outperforms the others, achieving an impressive accuracy of 97.01%, along with superior F1-score, recall, and precision metrics. This study provides important insights for improving cybersecurity frameworks by demonstrating a potential of cutting-edge ML algorithms in accurately identifying and blocking cyberthreats. Future directions include integrating hybrid models and real-time data processing to further improve detection efficacy in dynamic network environments.