User Modeling for Privacy-preserving Explanations in Group Recommendations

Research output: ThesisDissertation (TU Delft)

444 Downloads (Pure)


My thesis investigates what makes good explanations for group recommendations, considering the privacy concerns of group members. Let’s give an example. Have you ever been to lunch with other colleagues on a business trip? Do you recall how long it took you to pick a restaurant? In these situations, recommender systems could help people decide, e.g., where to go. Recommender systems are decision support systems helping users to identify one or more items that satisfy their requirements. Most often, recommender systems propose items to individual users. However, there are many scenarios where a group of users will consume a recommendation and need support for group decision-making. A group recommender system is a system that recommends items to groups of users collectively, given their preferences. An example is a system for suggesting places to visit to a group of colleagues traveling together. For example, think of a group decision regarding the next places to visit in a colleagues'/friends' group traveling. Explanations, for such recommendations, in this context, act as complementary information, describing how specific recommendations are generated to help the group make informed decisions on whether to follow or not follow recommendations. However, there are many types of information to include and many ways to formulate an explanation, and it is not clear which information should be shown in the explanation for a group. Besides, explanations for groups are different from explanations for single users in that they should consider the privacy aspect (e.g., people might be sensitive to disclosing some of their information in the group). In this chapter, I first introduce the motivation of this Ph.D. thesis of developing explanations for group recommendations/decisions context. To the best of my knowledge, this thesis is the first work that studied group explanations from the perspective when the privacy aspect is included. Then I list the research questions that guide my thesis to design explanations for groups and summarize the corresponding contributions. This includes studying what information to disclose and what not to disclose in a group explanation and what factors and how influence the decision of information disclosure in a group explanation, e.g., the group members' personality, the relationship between them, whether their opinion is aligned with the majority in the group or not. Finally, I present a list of publications carried out during this thesis.
Original languageEnglish
QualificationDoctor of Philosophy
Awarding Institution
  • Delft University of Technology
  • Houben, G.J.P.M., Supervisor
  • Tintarev, N., Supervisor
Award date1 Feb 2023
Electronic ISBNs978-94-6366-640-4
Publication statusPublished - 2023


  • Explanation
  • Personal information
  • Group recommendation
  • Information privacy


Dive into the research topics of 'User Modeling for Privacy-preserving Explanations in Group Recommendations'. Together they form a unique fingerprint.

Cite this