Smart malware detection on Android

摘要: Nowadays, because of its increased popularity, Android is target to a growing number attacks and malicious applications, with the purpose stealing private information consuming credit by subscribing premium services. Most current commercial antivirus solutions use static signatures for malware detection, which may fail detect different variants same zero-day attacks. In this paper, we present behavior-based, dynamic analysis security solution, called Malware Detection System, detecting both well-known malware. The proposed solution uses machine learning classifier in order differentiate between behaviors legitimate applications. addition, it application statistics determining reputation. final decision based on combination classifier's result includes unique extensive set data collectors, gather application-specific that describe behavior monitored application. We evaluated our applications obtained high accuracy 0.985. Our system able samples are not detected solutions. outperforms other similar running mobile devices. Copyright © 2015 John Wiley & Sons, Ltd.