Malware Classification to Strengthening Digital Resilience: Comparing SVM Kernel and Logistic Regression
Preet Singh, Taniya Hasija, Kr Ramkumar
Abstract
Malware is a type of software specifically designed to infiltrate digital systems by breaching the device’s security with a huge potential to cause serious damage to any organization by tampering with their sensitive data. The integration of Machine learning technologies capabilities to detect malware and mitigate attacks is an emerging trend in dealing with suspicious attacks and threats. This paper uses Support Vector Machine (SVM) and Logistic Regression (LR) to perform a supervised classification over the labelled dataset of malware, which is collected from Kaggle. This study also discovers the effectiveness of using different SVM kernels, including Linear, Polynomial and Radial Basis Function (RBF) kernels along with a detailed performance analysis. The best-performing SVM kernel is RBF with an accuracy of 98%. To make a sustainable digital space, steps are necessary to safeguard the risk of malware applications.