Android mobile security by detecting and classification of malware based on permissions using machine learning algorithms
P. Ravi Kiran Varma, Kotari Prudvi Raj, K. V. S. V. N. Raju · 2017 International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC) · 2017
Android occupies a major share in the mobile application market. Android mobiles have become an easy target for the attackers. The main reason is the user ignorance in the process of installing and usage of the apps. Android malware can be detected based on the permissions it requests from the user. Several machine learning algorithms are being used in the detection of android malware based on the list of permissions enabled for each app. This paper makes an attempt to study the performance of some of the machine learning algorithms, viz., naïve Bayes, J48, Random Forest, Multi-class classifier and Multi-layer perceptron. Google play store 2015 and 2016 app data are used for normal apps and standard malware data sets are used in the evaluation. Multi-class classifier was found to be outperforming the other algorithms in terms of classification accuracy. Naïve Bayes classifier has outperformed as far as model construction time is concerned.