Using virtual characters in personalized recommendations

This paper describes a virtual character placed in an educational environment with the purpose of motivating and involving students in the activities proposed. The character can communicate with users in natural-language; it can learn students' profiles and use this knowledge to recommend appropriate contents and activities. Three main elements compose the character's architecture: an AIML-based mechanism, which endows the character with its communication abilities; a profile management system, which collects and organizes students information; and a recommender system, which uses the student's profiles and behavioral patterns to make content recommendations. The character was placed in an educational environment and used in an experiment with 53 computer science students. The results of the experiment are presented here, as well as conclusions and directions for future work.