Android Adware Detection Using Machine Learning

Related Works

Many different approaches have been used to detect and classify android malware. Milosevic et al. (2017) present two machine learning methods to detect android malware using a machine learning model based on app permissions and app source code. These researchers achieved an F-score of 95.1% for source code based classification and a F-measure of 89% for permission based classification. Their approach used several classification algorithms such as C4.5 decision trees, random forest, and others. Their work differs from ours in that while their approach focuses on source code and app permissions, our work focuses on classification based on network traffic features.

Altaher & Barukab (2017) proposed an adaptive neuro-fuzzy inference system with fuzzy c-means clustering to improve the classification rate of android malware vs benign android apps. They used android permissions as features for classification. Their FCM clustering method determines the best number of clusters to improve classification accuracy. These researchers achieved a classification accuracy of 91% with a false positive rate of 0.5% and a false negative rate of 0.4%.