Malware Analysis in Cyber Security based on Deep Learning; Recognition and Classification
Mohamed Elalem, Tahani Jabir
Abstract
Cyber security in wireless communications can be an unwieldy subject, given the amount of malware that has been increasing rapidly in the last few years. This generates serious security problems for public agencies and the government institutions. In order to mitigate the influence of malware deployment and pervasiveness, new recent identification and classification algorithms for malware that adapt deep learning techniques are studied and figured out based on their features and behaviors. This study introduces a deep learning algorithm to identify different malware families. To implement the proposed approach, malware color-based images (RBG) are used directly. Then these malware images are identified and classified by considering the benefits of leveraged Convolutional Neural Networks (CNNs), which have the ability to automatically extract those features. A challenging malware classification experiment using the MaleVis dataset confirms that the adapted model outperforms better functional classification compared to the traditional machine learning models and achieves very good accuracy based on the MaleVis dataset.