Comparative Evaluation of Machine Learning and Deep Learning Models for Cyber Threat Detection

Cyber threats are evolving rapidly, posing significant risks to critical infrastructure, organizations, and individuals.This study investigates the effectiveness of machine learning which are random forest and gradient boosting including deep learning which are long shortterm memory and convolutional neural network for cyber threat detection.This research uses dataset from the University of New South Wales namely UNSW-NB15.The methodology involves data analysis, data preprocessing, features engineering, models training and performance evaluation.A comprehensive set of preprocessing techniques to improve data quality and model performance is included.Models are evaluated using accuracy and training time.Results show random forest outperforms the others with the highest accuracy (80%) and fastest training time (55 seconds), demonstrating its suitability for real-time applications.