Android Malware analysis using Java and SVM

Abstract

The evolution of operating systems and the rising number of smartphones in the market has been grabbing a lot of attention in recent years. The capabilities and functionalities of android mobile devices have been recently expanded by a large number of third-party applications. Although technological advancements and innovative applications attract a lot of businessmen and scientists alike, it also entices malicious attackers and hackers. Applications installed in a device present a way for the attackers to breach the security of the system. With the growth of technology, there is a real need to understand the threats of malware and the process to nd them and protect the system from attacks. Application developers often ask unnecessary and wrong permissions such as misspelled, non-existent, deprecated, or protected ones. The data retrieved from such permissions are prone to malware attacks. Although the android system uses coarse-grained permissions from the user to warn about sensitive information access by applications, users usually have less knowledge of or control over how their privacy-sensitive data is used. The le operations are done by the application in the user's device has a severe impact if used for malicious purpose. This study is focused on analyzing multiple android applications and attempts to identify the characteristics of this application code with static and dynamic analysis using the Java Weka tool and classify them using the machine learning algorithm.