A Hybrid Approach to Ransomware Detection Using Convolutional Neural Networks and Random Forests on Network Traffic Patterns
Lala Woo, Shi Tan, Tsingyi Tan, Gua Lao · 2024
Ransomware attacks have rapidly escalated into one of the most pressing cybersecurity threats, causing severe financial and operational damage to both individuals and organizations. Traditional detection methods often fall short of identifying evolving ransomware strains, as they rely heavily on static signatures and predefined rules. A novel approach is presented that integrates Convolutional Neural Networks (CNNs) and Random Forests to enhance detection accuracy through network traffic analysis. The CNN component automatically extracts complex patterns from network flows, while the Random Forest classifier ensures robust classification of ransomware-related anomalies. This hybrid model significantly improves detection performance, reducing false positives and adapting to new ransomware variants in real-time. The findings demonstrate the effectiveness of machine learning in automating ransomware detection and offer a scalable solution that can be applied to diverse network environments.