Android Adware Detection Model Based on Machine Learning Techniques

In recent years, the popularity and use of smartphones have increased which make it the main medium of communication. Attackers are constantly monitoring smartphones to get secret data and information from them through various attacks of malware, one of them is Advertising Software (Adware). Security specialists and researchers are working on the effective detection of android malware using Machine Learning (ML) techniques. In this paper, a new detection model is proposed to protect smart devices from adware attacks via monitoring the network traffic. Several data pre-processing, feature selection techniques, and ML algorithms are used to detect adware samples in is presented dataset. Then, a comparison was made between the ML classifiers, Random Forest (RF), k-Nearest Neighbors (k-NN), Decision Tree (DT), Multi-Layer Perceptron (MLP), XGBoost (XGB), and Logistic Regression (LR) via seven performance metrics to determine the best for adware detection. The proposed method showed the best detection accuracy are (98.66%), (98.10%) and (98.05%) for (DT), (XGB) and (k-NN) respectively.