A procedure of how to conduct research in transparent mobile recommendations

Abstract

The information overflow of today's information society can be overcome by the usage of recommender systems. Due to the fact that most recommender systems act as black boxes, trust in a system decrease, especially when a recommendation failed. Recommender systems usually don't offer any insight into the systems logic and cannot be questioned as it is normal for a recommendation process between humans. Transparency, which is about explaining to the user why a recommendation is made, supports the user in a way of understanding the reasoning behind a recommendation. Within a mobile environment, it is possible to address the user more individualized but transparency needs a completely different way of visualization and interaction. The paper in hand aims at setting up a process model on how to address transparency in mobile recommendations and therefore introduce into a complex new area of research, recommender systems didn't address in the past. © 2008 International Federation for Information Processing.