Malware Detection Using Machine Learning on Edge Devices

Hackers’ primary attack vectors against an IoT network are IoT nodes and edge devices. As a result, developing practical and reliable methods for identifying malicious behavior of the nodes is a critical step toward safeguarding the entire network. Recent research on this topic has been limited to less accurate Machine Learning models and limited analysis of optimizers and activation functions. Additionally, the impact of false negatives in malware detection makes it more challenging. To address this issue, this paper uses a Deep Neural Network (DNN) to develop different models for analyzing malware detection with various optimizers and activation functions. A benchmark dataset, Drebin is used to train and evaluate the proposed model. The results demonstrate that the proposed DNN model has a relatively good accuracy (99.2%) and F1-score (98.88%) in comparison with the existing techniques.