App Doppelganger Detection via Artificial Neural Network

Abstract

Detecting repackaged applications is one of the most pressing issues in the Android universe. Several attackers typically find a good program, alter or insert malicious code into it, repackage it, and sell it in the online business sectors. They use code obfuscation tactics to discourage app cloning and repackaging. Good code protection is another tool for identifying repackaged Android applications that we suggest. The approach examines the similarity of Android applications based on the method call specifics of component classes that receive inferred intents. We offer a counting-based method for efficiently filtering the noisy layout that can lead to deviation. It can distinguish between the program with absolute and partial level similarities. Another benefit of our approach is precision. We suggest an artificial neural network technique for filtering the noisy layout that can induce variance. It can detect absolute and partial application-level similarities. We assess the clone recognition approach on a constant informational index and check the Review and Exactness, which is critical. Moreover, our discoveries show that our technique is inconceivably excellent and precise in identifying different kinds of clones to decide the level of similitude in Android applications.