Analysis of Machine Learning Classifier in Android Malware Detection Through Opcode

Abstract

As the development of technology in Android devices continue to rise, more mobile applications were introduced to make daily task much easier. The mobile applications has also used as platform for users to do social networking, online banking, online shopping, web browsing, and many more activities. Most of these applications requires user to provide private credentials, which could potentially be sniffed by cybercriminal and exploited. Various detection techniques have been introduced in mitigating mobile malware, yet the malware author has its own method to overcome the detection method. Therefore, an enhanced mobile malware detection technique is needed to defend smartphone users against malicious threats. Based on this reason, this paper explores the mobile malware detection through opcode analysis and proposed the optimum machine learning classifier method. Several machine learning classifier methods are selected based on the previous research, and their performance are analyzed. The machine learning classifier methods are evaluated based on the contribution it made to improve the True Positive Rate (TPR), False Positive Rate (FPR) in classifying benign and malicious mobile malware application.