SVM Based Approach for Detecting Malware in Android Using Machine Learning

Abstract

At present, Android has gotten one of the most well- known operating system for cell phones on account of various mobile apps it supports. Nonetheless, the downloaded malevolent Android apps (malware) from outsider markets have altogether threatened privacy and security of the clients. The large portion of malwares stay undetected because of the absence of effective and precise malware recognition methods. In this contribution, we discuss a SVM based methodology to identify the malware for Android system, that incorporates both hazardous authorization mixes and defenceless API calls and utilize them as highlights in the machine learning approach. In order to test the performance of presented methodology, broad analyses have been organized, that demonstrated that proposed scheme can recognize pernicious Android apps viably and effectively. By making use of experimental verification, we prove that SVM beats rest of the machine learning classifiers.