SVM-Based Machine Learning Approach for Mobile Botnet Detection

Abstract

Botnets pose a significant threat to global Internet security. A new pattern is evolving to shift botnets from conventional desktop to mobile platforms with continued sophistication and durability surroundings that move. To lessen the harm that mobile botnets pose, just like in the desktop environment, detection is crucial. Recognizing patterns of unusual behaviour is one of the many methods used to find these botnets, and it produces the best and most often findings. Analysing the running characteristics of this kind of application is one technique to spot similar tendencies in the mobile botnet area. This article examines host-based and anomaly-based methods for detecting mobile botnets. The suggested method extracts statistical characteristics of aberrant behaviour from system calls and utilises machine learning methods to identify them. In realistic settings, the effectiveness of the employed strategy can be tested. Excellent outcomes were attained by the suggested strategy, which included a low false positive rate and a high actual detection rate.