Anomaly Detection in Network Traffic Using Machine Learning and Deep Learning Models

Nowadays, cybersecurity is of utmost importance as most organizations integrate technology into their internal management systems. This integration, however, often leads to malicious attacks on these systems for various purposes, such as data theft, system disruption, or discrediting the organization. These challenges compel organizations to protect their data and systems from such malicious actors. Cyberattacks, which can severely damage organizations, are frequently executed through network intrusions. As a result, organizations need to have a network security system that can respond to threats immediately. This paper discusses an overview of cybersecurity threat prevention and network security systems, focusing on anomaly detection in network traffic using Random Forest (RF), a supervised machine learning algorithm, and Convolutional Neural Networks (CNN), a deep learning technique. Additionally, it compares the performance of these two models using metrics such as Accuracy, Precision, Recall, and F1-Score to analyze their capabilities in detecting anomalies. The paper presents a comprehensive process encompassing data preparation, model training, and model evaluation. This serves as a guideline for developing robust network anomaly detection systems, ultimately enhancing organizational confidence in strengthening cybersecurity in the digital age.