Measuring the Learnability of Interactive Systems Using a Petri Net Based Approach

A learnable system allows a user to know how to perform correctly any task of the system after having executed it a few times in the past. In this paper, we propose an approach to measure the learnability of interactive systems during their daily use. We rely on recording in a user log the user actions that take place during a run of the system and on replaying them over the system interaction models, which describe the expected ways of executing system tasks. Our approach identifies deviations between the interaction models and the user log and assesses their weight through a fitness value. By measuring the rate of the fitness value for subsequent executions of the system we are able not only to understand if the system is learnable with respect to its tasks, but also to quantify its degree of learnability over time and to identify potential learning issues.