Malware Classification using Static Analysis Approaches

Abstract

Malware threats are becoming increasingly complex, thereby posing greater challenges to effective mitigation efforts. It has become more essential than ever to address the malware, as it poses significant threats to individuals, organizations, and governments worldwide. Therefore, effective and more advanced malware classification techniques are necessary to address these malware threats. This proposed study presents an advanced approach to malware classification using static analysis which examines files without executing them. A structured framework was developed for systematic classification which involves gathering raw malware samples from various sources. Raw malware samples were deconstructed and information, such as the frequency of the opcode and the size of the section were collected. Experiments were carried out to assess the effectiveness of several classifiers in terms of accuracy, precision, recall, and F1 score across distinct malware classifications. Random Forest emerged as the best model in the examined dataset, with an accuracy of over 85.0%. These results demonstrate the effectiveness of Random Forest based on extracted datasets. The proposed research focuses on malware detection and classification, thereby enhancing cybersecurity in modern computing environments.