Predictive Menu Selection on a Mobile Phone
Robert Bridle, Eric C. McCreath · 2005
Abstract. A mobile phone is often characterised by its limited user interface. We consider applying a machine learning approach to improve the usability of a mobile phone’s interface. We present an approach that predicts a user’s menu selection on a mobile phone. The learning approach consists of a highly restricted set of hypotheses. A restricted set of hypotheses allows both the learner to operate efficiently within the limited hardware resources of a mobile phone and learn concepts quickly from a small training set. We compare our approach with alternative menu prediction approaches in a simulation. Also a modified Nokia TM Series 60 mobile phone address book application was developed to incorporate our menu prediction approach. A user evaluation of this implementation shows that our predictive menu selection approach has the potential to reduce the number of key presses a user makes. 1