Supporting Serendipity through Interactive Recommender Systems in Higher Education

Serendipity is defined as the surprising discovery of useful information or other valuable things. In recommender systems research, serendipity has become an essential experiential goal. However, relevant to Human-Computer Interaction, the question of how the user interfaces of recommender systems could facilitate serendipity has received little attention. This work investigates how recommender system-facilitated serendipity can be applied to research article recommendation processes in the context of higher education. In particular, this work investigates the use of recommender system applications in developing countries as most studies in developing countries have focused solely on implementation, rather than user experiences. This dissertation describes the design and development of several user interfaces for recommender systems in an attempt to improve our understanding of serendipity facilitation with the help of user interfaces. By studying these systems in a developing country, this dissertation contrasts the study of recommender systems in developed countries, examining the contextual and cultural challenges associated with the application of recommender systems.

This dissertation consists of five empirical user studies and a literature review article, contributing novel user interface designs, open-source software, and empirical analyses of user experiences related to recommender systems in a Pakistani higher education institution. The fortunate discoveries of recommendations are studied in the context of exploring research articles with the help of a recommender system. This dissertation covers both constructive and experimental research. The articles included in this dissertation present original research experimenting with different user interface designs in recommender systems facilitating serendipity, discuss stakeholder requirements, assess user experiences with recommended articles, and present a study on task load analysis of recommender systems. The key findings of this research are that serendipity of recommendations can be facilitated to users with the user interface. Recommender systems can become an instrumental technology in the higher education research and developing countries can benefit from recommender systems applications in higher education institutions.