Facilitating Discovery of Books Through Proactive User Cues

Search and recommendations drive the majority of sales at KOBO. It is therefore critical to continually improve and tune these algorithms. Our project will explore several questions dealing with integrating information provided by users in order to optimize search and recommendation algorithms: how to integrate user-generated tags in order to enhance search (rather than simply depending on product metadata), how to elicit user preferences when they join the service (when we have no existing data to work with), and how to present reasons for recommendations to users, as well as how to integrate user feedback on these recommendations.

Faculty Supervisor:

Eugene Fiume

Student:

Sasa Milic

Partner:

Kobo Inc.

Discipline:

Computer science

Sector:

Information and communications technologies

University:

University of Toronto