GR/BT/MT Adapting Heuristic Evaluation Methods for Early-Stage Recommender Systems User Interface Design

Recommender systems recommend items to an active user based on information about users and items. The evaluation of recommender systems is often based on offline datasets, but testing the potential benefits of an approach from the users' perspective is also important [1,2]. For user-centered evaluation, it is often required to already have a high-fidelity prototype with which users can interact. However, it is desirable to test user interface design in the very early phases of system design as well. For this general task, Heuristic Evaluation (HE) has been developed in Human-Computer Interaction (HCI) as an analytic method for finding usability problems in the early phase of user interface design, see [3] for an overview and [4] for a specific approach.

The goal of this project is to adapt Heuristic Evaluation to the specific requirements of early-stage user interface design for recommender systems and develop specific guidelines. For example, user interface options for preference elicitation, control of the recommendation process, and feedback on recommended items should be explicitly considered. The methodology and guidelines can then be tested in a few case studies. This topic is suitable for a Guided Resarch (IN2169) or a Bachelor's or Master's Thesis in Informatics or related study programs, with the scope adapted accordingly.

Prerequisites are high motivation and knowledge or at least interest in HCI methods. Please send your application (brief CV, transcript of records and short motivation statement) to Wolfgang Wörndl ( until March 14th. (Decision and possible start soon after this date, a Guided Research has to be registration by the first week of the new semester.)

[1] Gunawardana, A., Shani, G., & Yogev, S. (2012). Evaluating recommender systems. In Recommender systems handbook (pp. 547-601). New York, NY: Springer US.

[2] Zangerle, E., & Bauer, C. (2022). Evaluating recommender systems: survey and framework. ACM Computing Surveys, 55(8), 1-38.