Harmony in melody : a mediator concept for song recommendation

Abstract

This thesis discusses the adaptation of dialogue systems based on user personality. With the increasing integration of computers into people's private and professional lives the importance of effectively navigating computer systems seems to be a valid concern. As not all user interfaces can be used intuitively a bridge to close the gap between the user's expectations and the system's capabilities seems necessary. Therefore, this thesis offers the conceptualization of an adaptive dialogue system that acts as a bridge between a system and a user. For this thesis, the online music streaming platform Spotify will be used as the system's domain in question. Further, the proposed system will adapt to a user's personality. Therefore, personality traits are discussed and two traits will be chosen as axes for adaptation. Based on these, four policies will be designed to match each combination possible. To test whether adaptation based on personality makes a difference to the user, a Wizard of Oz study will be conducted. Therefore, users will be asked to create a playlist using the dialogue system. The study will then be evaluated and discussed. Interactions with the dialogue system will then be annotated to create a database as a result of this thesis.