Developing a novel methodology for virtual machine introspection to classify unknown malware functions

Abstract

In recent years, cloud service security is a crucial task because of several vulnerabilities in Virtual Machine such as malicious activities, viruses, and errors. Therefore detecting malicious activity is essential to improve the security of the cloud and VM. There are many existing techniques are developed to identify malicious attacks but still having the issues of less accuracy of detecting attacks, high false prediction rate and error also the main problem is the complexity to detect malware attacks because of large files. So, this current research proposed a new Adversarial-based Generative Model with African Buffalo (AGM-AB) technique to classify unknown malware functions presented in the VM. Also, AB fitness is initializing in AGM for enhancing the performance of feature extraction and classification. In addition, the developed AGM-AB technique categorizes executable files of benign and malware also improve the accuracy of malware detection. Furthermore, launch the unknown malware in developed technique for validating the efficiency of classification in AGM-AB technique. Thus the developed AGM-AB technique is implemented in python, and the performance metrics are calculated such as accuracy, AUC, False Positive Rate (FPR), recall value, precision, and F-measure.