GADroid: A framework for Malware Detection from Android by using Genetic Algorithm as Feature Selection approach
Abstract
Android has gained its popularity due to its open-nature and number of free apps present in its official play-store. To see the exponential growth of Android-based devices, cybercriminals are developing malware-infected apps and publishing them into different promise repositories for users. These malware-infected apps are of great threat to the user’s privacy. For this reason, it becomes necessary to develop a malware detection model. Therefore, in this research paper to develop a malware effective model, we select features by using a genetic algorithm as feature selection approach. By considering selected features as input, we trained it with the help of Deep Neural Network (DNN) machine-learning algorithm. The experimental results proves that the model developed by using 5,60,142 distinct Android apps, which belongs to thirty distinct categories of Android apps, can detect 98.6% malware from real-world apps.
Keywords: Smartphone, Permissions model, API calls, Deep Neural Network (DNN), Feature selection, Intrusion detection, Cyber security, Android apps.