DATA MINING BASED MALICIOUS APPLICATION DETECTION OF ANDROID

One of the most popular mobile phone platforms is an Android mobile phone platform owned by Google. The Android platform is open source to allow the developers to develop the full future application of the mobile operating system. Nowadays, malicious applications have been expanding in scale as an Android system. In this paper a data mining aided approach to detect malware applications in Android applications is presented. This approach capture the instant attracts that cannot be conclusively identified in past work. Static detection is one of the popular methods based on permissions detection of maliciousness in all the way through AndroidManifest.xml by classifiers. This paper suggests implementing a malicious application identify tool called Androidspy. Initially observe the relationship among system functions, sensitive permissions, and interface of responsive programming. Then, examine the system function grouping that has been clarifying the application behavior and characteristic vector. Following on the characteristic vectors, finding malicious android applications used to be naive Bayesian, function decision algorithm, methodologies of j48decision tree. Androidspy is real-world applications as well as test sample programs. The test sample result confirms that Androidspy can be enhanced to detect malicious applications by using the system function group estimated with the previous work.