Android Malware Detection System using Machine Learning

Abstract

During the past year until now, the amount of malware targeting Android operating system has been rising dramatically. Therefore, Android malware detection are required to detect the malware before getting more serious. The static analysis examines the full code of application meticulously while dynamic analysis identifies the malware applications by monitoring it behaviors. This study proposed a malware detection system by using machine learning approach and aims to detect malware that has attacked Android operating system. In this research paper, the Android malware detection system are trained using five types of classifiers meanwhile WEKA is used for simulation process. The dataset used contains 10k of malware and 10k of benign. The outcomes presented Random Forest classifiers attained highest accuracy result, 89.36% compared to Naïve Bayes which 89.2%. TPR is viewed as detection rate which precisely predicted malware process while FPR is choosed as detection rate which inaccurately predicted normal as malware. To evaluate detection exactness which is good or bad, the area under the curve (AUC) have been applied through this study. The results show that Naïve Bayes has the lowest model complexity as it uses minimal time to build the model. Hence, it can be concluded that achieving reasonable accuracy and effectiveness in classifying unknown malware helps to determine the performance of the classifiers.