Predicting Future User Actions by Observing Unmodified Applications

摘要

Intelligent user interfaces often rely on modified applica-tions and detailed application models. Such modifications and models are expensive to build and maintain. We pro-pose to automatically model the use of unmodified applica-tions to solve this problems. We observe a suer's interactions with the application's interface and from these observation infer a state space which the user navigates and the stochas-tic policy he or she follows. ONISI, the algorithm presented here, builds this state space implicitly and on-line, and uses it to predict future user actions. Trials with real users show that this algorithm predicts the next user action significantly better than another current algorithm.