Context based adaptation of application icons in mobile computing devices

Finding the desired application among a huge number of installed applications in a mobile computing device, like a smartphone or tablet, is problematic for some users. In this paper we propose a novel method to adapt the display of application icons in mobile devices based on context. In our method, the icons that are predicted to be used are displayed prominently. The prediction is based on the current application context being invoked, incorporating learning from past user actions associated with the current application. The system predicts the next applications, whose icons are displayed in a way to attract the user´s attention and thus reduce the access time. This allows the user to perform predicted actions in minimal steps and also reduce the clutter on the screen. We use machine learning algorithms to predict the best possible screen configuration with respect to the context or scenario in which user is present, and can be generalized for any user interface (UI) related aspect rather than just the positioning of icons. We study the average response times for a number of users to access certain applications randomly on a phone, and on that basis predict the utility in terms of time saved by adapting the UI in this way. Our approach assists the users to more quickly access the desired application.