Detecting Ransomware using GURLS

Abstract

Ransomware is a malware, which upon execution scrambles the framework and it denies the client from accessing the data until the point when a payoff sum is not met from the victim. Recently, this kind of malware has shown a massive growth and had affected nearly 100 nations around the globe. In this paper we propose GURLS (Grand Unified Regularized Least Square) based approach to detect ransomware and classify it into different categories. The features used for training and testing are application programming interface (API) invocations and strings. This paper compares the performance of each of these features for classification and the effectiveness of RBF Kernel. The results obtained shows that using RBF kernel gives better accuracy.