Supervised Machine Learning Framework for Fraud and Malware Detection in Android Apps

Abstract

At present, everyone is dependent upon its Smartphone for banking, communication, business,
gaming and many more functionalities. But,Ransomware is one of today's most severe Internet
security challenges and also Android applications also effective by the various types of Trojan attacks
respectively. Indeed, most Internet issues, including spam e-mails and denial of service attacks, are
triggered by malware and android applications also facing this issue. In many words, Smartphone’s
that are infected by Ransomware are also networked into botnets, and often assaults are performed on
hostile, assaulting networks. From untrusted internet sites may be likely to contribute to
maladministration. These executables are changed intelligently to circumvent antivirus specifications
by anomalous users. In this article, an improved identification approach for harmful executables is
suggested by evaluating Portable Executable (PE) executable files and utilizing an extraction process
for support vector machine (SVM)classification. We also learned a supervised binary classifier using
these features from regular and malicious PE data on Android applications. We have checked our
system on a comprehensive publicly accessible dataset and obtained a rating maximum accuracy
compared to the state of art approaches respectively.