A Cutting-Edge Deep Learning Framework for Streamlined Malware Detection in Cybersecurity Utilizing Long Short-Term Memory (LSTM)

Abstract

Malware is a malicious piece of code that has been causing security issues to information systems and networks across the globe. The entire cyberspace is suffering security issues due to malware. The traditional approaches based on heuristics could provide specific mechanisms to protect information systems from malware. However, the ever-increasing malware patterns made it very difficult to detect new malware with the help of heuristics-based methods. The emergence of Artificial Intelligence (AI) has paved the way for learning-based approaches that could provide better performance in malware detection. However, Enhancements are required to the current deep learning techniques utilized in malware detection in addition to updating trading data used for the supervised learning process. Our proposal in this study was a deep learning framework for efficient malware identification in cybersecurity with the use of Long Short Term Memory (LSTM). Our approach, which we called Learning-Based Malware Detection (LBMD), which exploits the efficient dataset created as part of the framework and uses the LSTM model appropriately to detect malware efficiently. Our empirical study with the dataset created reveals that the proposed algorithm outperforms many existing malware detection models like KNN, XGBC, and baseline CNN with the highest accuracy of 98.79%.