To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
Cloud Security Based Malware and Ransomware Detection Model Using Supervised Machine Learning Techniques |
|
Author |
Raed Alharthi |
| Citation |
Vol. 25 No. 4 pp. 218-226 |
|
Abstract |
Cloud computing offers several kinds of computer services via the Internet. Instead of utilizing their local storage, many consumers and enterprises are adopting the cloud to store their data remotely in data centers. This allows access to data from any device, at any time, and from any location. Protecting privacy and dealing with malware threats are two of the main issues facing cloud computing. Ransomware attacks have seen hackers take important data hostage in exchange for financial gain. The encrypted nature of ransomware-infected files makes it challenging to retrieve original data from them without special keys. The accuracy of malware detection has been the subject of several research, but the privacy protection of cloud tenants has not received enough attention. This research presents a novel feature extraction, selection, and detection approach for cloud-based semi-supervised transfer learning (SSTL) malware and ransomware detection. CatBoost classifiers are intended to identify malware and ransomware files to protect tenants' privacy in public cloud environments. Temporal Convolution Networks (TCNs) are used in this phase to calculate features, feature action kinds, and path scores. First, each property's frequency is computed. The Sand Cat Swarm Optimization technique is utilized to select features based on path scores and action states once the frequency computation is complete. Following feature selection, we categorize malware and ransomware attacks using the CatBoost algorithm. In terms of detecting and categorizing attacks, the CatBoost method performs better than other machine learning classifiers. |
|
Keywords |
Malware detection; Ransomware detection; Temporal Convolution Networks; Machine learning. |
|
URL |
