Malware Detection in Android OS using Machine Learning Techniques
Maad M. Mıjwıl
Abstract
Malware is a software that is created to distort or obstruct computer or mobile applications, gather sensitive information or execute malicious actions. These malicious activities include increasing access through personal information, stealing this valuable information from the system, spying on a user’s activity, and displaying unwanted ads. Nowadays, mobile devices have become an essential part of our times, therefore we always need active algorithms for malware detection. In this paper, supervised machine learning techniques (SMLTs): Random Forest (RF), support vector machine (SVM), Naive Bayes (NB) and decision tree (ID3) are applied in the detection of malware on Android OS and their performances have been compared. These techniques rely on Java APIs as well as the permissions required by employment as features to generalize their behavior and differentiate whether it is benign or malicious. The experimentation of results proves that RF has the highest performance with an accuracy rate of 96.2%.