Hybrid intelligence: integrating machine learning and nature-inspired algorithms for malware detection

Hybrid intelligence: integrating machine learning and nature-inspired algorithms for malware detection

연구 분야: Safety

학회: Journal of Computer Virology and Hacking Techniques

This paper presents an innovative approach of cyber threats detection by a hybrid learning method using machine learning and nature inspired algorithms. Various kind of malware attacks such as Trojan, spyware, ransomware are increasing day-by-day. Ransomware attack is the greatest concern of recent days as it causes both information loss and financial loss. If we can detect such types of attacks from the network traffic, it may result in a successful prevention of these attacks. We have considered the CICMalMem-2022 dataset and executed our proposed hybrid attack detection model on this dataset. Firstly, we have executed a Decision Tree and Random Forest model on the dataset. Next, we have introduced a new ensemble learning techniques using several Nature Inspired Algorithms. The main objective of this proposed model is to optimize the feature set and then perform the attack detection task based on the optimized features. The experimental results harness the acceptance probability of the proposed models for such kind of optimization and attack detection tasks. We have also shown the comparisons of the proposed works with other existing Machine Learning models.