Baseline for Performance Prediction of Android Applications

Millions of applications are being developed around the world with a promise of increasing our productivity, teaching us new skills, improving our health and providing us with entertainment. In order to cater to the users needs, ever more intricate and advanced applications are being developed and with more sophisticated solutions, an increase in resource expenditure is likely to occur. Smartphones have limited resources in terms of both computational performance and battery capacity and most developers today tries to optimise their applications in order to provide the best user experience with as low a resource cost as possible. In this paper, large amounts of performance data has been gathered from different Android applications and games on different devices in order to establish a baseline which in turn can be used for training of machine learning algorithms for usage in automatic testing and real-time performance assessment solutions.