Automatic malware categorization based on K-means clustering technique

The android operating system is a popular operating system for mobile phone applications. This is also known as an open-source operating system so that the developers can easily update and add new features to it. However, it poses significant challenges related to malicious attacks or cyberattacks because of its open system design philosophy. Nowadays, the number of malware applications is increasing rapidly and proportionally with safe android applications. As a result, it has become very challenging to identify their behaviors or signatures or categorizes them to implement protection in the android system. In this research work, we propose an automated system for malware categorization using the K-means clustering method that automatically chooses the cluster number. In our method, we have categorized malware into an optimum number of different cluster families by using a real-time malware dataset. We also compare our automated model with the traditional cluster selection technique with Elbow and Silhouette method. Experimental results demonstrate that our model determines the optimal cluster number with less user intervention for malware categorization.