[Paper Review] Finding the Right Set of Users: Generalized Constraints for Group Recommendations

Kostas Stefanidis, Evaggelia Pitoura|arXiv (Cornell University)|Feb 26, 2013

Recommender Systems and Techniques14 references3 citations

TL;DR

This paper introduces a generalized constraint model for group recommendation that selects optimal user groups based on both item relevance and interpersonal compatibility constraints. It proposes a greedy algorithm that iteratively adds users who maximize group consensus while satisfying user-defined and system-level constraints, ensuring high collective satisfaction and compatibility in group formation.

ABSTRACT

Recently, group recommendations have attracted considerable attention. Rather than recommending items to individual users, group recommenders recommend items to groups of users. In this position paper, we introduce the problem of forming an appropriate group of users to recommend an item when constraints apply to the members of the group. We present a formal model of the problem and an algorithm for its solution. Finally, we identify several directions for future work.

Motivation & Objective

• To address the gap in group recommendation systems that do not account for user preferences regarding group composition.
• To formalize constraints on user-to-item, user-to-group, and group-to-group compatibility for group formation.
• To develop an algorithm that selects a group of users who are both interested in the recommended item and compatible with each other based on constraints.
• To maximize collective satisfaction by balancing item relevance and interpersonal compatibility in group selection.

Proposed method

• The paper models user preferences as constraints on item attributes (user-to-item) and on group member attributes (user-to-group and group-to-group).
• It introduces a greedy algorithm that selects users one at a time, prioritizing those who, when added, maximize the number of satisfied constraints or minimize the number of violated constraints.
• The algorithm evaluates potential users based on the impact of their addition on constraint satisfaction, using a scoring mechanism that considers both individual and collective constraint fulfillment.
• The method supports flexible constraint types, including value-based preferences (e.g., age, education) and cardinality constraints (e.g., minimum number of users over 40).
• It integrates both user-level and system-level constraints, such as a company’s preference for middle-aged or senior travelers.
• The algorithm terminates when the desired group size is reached, ensuring the final group satisfies the highest possible number of constraints.

Experimental results

Research questions

• RQ1 How can group recommendation systems be extended to consider not only item relevance but also interpersonal compatibility among group members?
• RQ2 What formal model can represent diverse constraints on group composition, including user preferences for other group members?
• RQ3 How can an efficient algorithm be designed to form a group that satisfies the maximum number of constraints while maximizing collective interest in the recommended item?
• RQ4 What trade-offs arise when balancing item relevance and group compatibility, especially when no group satisfies all constraints?

Key findings

• The proposed greedy algorithm successfully forms user groups that satisfy a high number of user- and system-level constraints while maximizing collective interest in the recommended item.
• The algorithm demonstrates effectiveness in selecting group members who not only have high individual interest in the item but also fulfill interpersonal compatibility constraints.
• The method supports flexible constraint modeling, including value-based preferences (e.g., age, education) and cardinality constraints (e.g., minimum number of users over 40).
• The approach enables the recommendation of items to groups even when no group satisfies all constraints, by prioritizing groups that violate the fewest constraints.
• The algorithm is scalable and practical for real-world applications such as travel packages and course recommendations, where group compatibility is critical.