Wisdom of the Crowd: Incorporating Social Influence in Recommendation Models

Recommendation systems have received considerable attention recently. However, most research has been focused on improving the performance of collaborative filtering (CF) techniques. Social networks, indispensably, provide us extra information on people’s preferences, and should be considered and deployed to improve the quality of recommendations. In this paper, we propose two recommendation models, for individuals and for groups respectively, based on social contagion and social influence network theory. In the recommendation model for individuals, we improve the result of collaborative filtering prediction with social contagion outcome, which simulates the result of information cascade in the decision-making process. In the recommendation model for groups, we apply social influence network theory to take interpersonal influence into account to form a settled pattern of disagreement, and then aggregate opinions of group members. By introducing the concept of susceptibility and interpersonal influence, the settled rating results are flexible, and inclined to members whose ratings are “essential”.

💡 Research Summary

The paper addresses a notable gap in modern recommender systems: the under‑utilization of social network information beyond simple similarity measures. While collaborative filtering (CF) remains the dominant technique, it treats users as isolated rating vectors and ignores the rich dynamics of social influence that shape real‑world decision making. To bridge this gap, the authors propose two complementary models—one for individual recommendation and one for group recommendation—both grounded in theories of social contagion and influence networks.

Individual Recommendation Model
The core idea is to augment the standard CF prediction (r_{ui}^{CF}) with a contagion‑derived term (s_{ui}). Each user (u) possesses a susceptibility parameter (\theta_u) that captures how prone they are to being swayed by peers. For every neighbor (v) in the social graph, an influence weight (w_{uv}) quantifies the strength of the relationship. The probability that a rating cascades from (v) to (u) is modeled as (p_{uv}= \theta_u \cdot w_{uv}). When enough neighbors have expressed high ratings for an item, a cascade is triggered and (s_{ui}) becomes a weighted average of those neighbor ratings. The final prediction is a convex combination:

\