Application of Trace-Based Subjective Logic to User Preferences Modeling.

A good way to help users make decisions in an interactive application consists in suggesting choices in accordance with their preferences. This decision problem faces challenging tasks, mainly in choosing a good solution that satisfies users and reaches the defined goal. Classical decision methods take into account the goal, but not all the obtained decisions can satisfy users’ preferences. The originality of our explorative research is to associate Subjective Logic (SL) to system’s traces (historical information) in order to model the user preferences that improve the decision process. Following JOsang, SL provides a suitable framework for modeling and formally describing users’ preferences. We propose to connect data collected in past executions, called traces, to the user intuition in order to support subjective reasoning. Based on this result, we can choose a reasonable decision according to users’ preferences. A Tamagotchi system will be presented to validate our result.