Enhancing 5G Security Using Hybrid Learning Approach for Malware Detection and Classification
E. Murugavalli, K. Rajeswari, V. Vinoth Thyagarajan, P. Shruti, K. R. Hemalatha · Advances in information security, privacy, and ethics book series · 2024
As the security challenges are inherent to 5G networks, the threat of malware continues to rise. Due to the high-speed nature of 5G, there is a need for automated malware detection and classification methods. This chapter presents a hybrid model for malware classification that combines machine learning and deep learning techniques. This model leverages the power of feature extraction using the InceptionResNetV2 deep learning model pre-trained on ImageNet. The features extracted from malware images are fed into a random forest classifier for the final classification task. The experimental evaluation conducted on a dataset comprising 31 malware families achieves an accuracy of 94.15%. Furthermore, the predicted labels closely align with the ground truth labels across various malware families, showcasing the model's ability to capture the intricate patterns and characteristics of diverse malware strains.