An IoT enabled Malware Identification Mechanism over Digital Documents using Predictive Learning Scheme

Abstract

The ability to identify malicious software is crucial to ensuring the safety of computer networks. Nevertheless, signature-based technologies now in use are inadequate for identifying zero-day assaults and polymorphic infections. That's why we need detection methods that use machine learning. The proliferation and sophistication of malware threats have elevated the problem of automated malware identification to the forefront of network security discussions. Manually analyzing all malware in a programme using conventional malware detection techniques is laborious and resource-intensive. The proliferation of the internet has led to a meteoric rise in the number of people using IoT devices. Malware assaults are growing more common as the storage capacity of IoT devices grows; as a result, detecting malware in IoT devices has become a pressing concern. For identifying malware in IoT devices, we present an ensemble categorization approach based on deep learning. For the identification of malware, we employ ANN and LSTM outputs. Our suggested technique achieves an average accuracy of 99.49% on standard datasets, which is higher than the accuracy of state-of-the-art methods.