Exploiting Implicit Influence From Information Propagation for Social Recommendation.

Area of Science:

• Computer Science
• Information Science
• Social Computing

Background:

• Social recommender systems are widely studied.
• User ratings and reviews on social media implicitly influence future user ratings.
• Implicit influence propagation among users who rated the same items significantly impacts recommendations but is understudied.

Purpose of the Study:

• To propose an information propagation-based social recommendation method (SoInp).
• To model implicit user influence from the perspective of information propagation, inferred from ratings on the same items.
• To investigate the effect of implicit user influence and incorporate it with explicit trust information into recommender systems.

Main Methods:

• Developed the SoInp method to model implicit user influence via information propagation.
• Inferred implicit influence from users' ratings on shared items.
• Integrated implicit user influence and explicit trust information within a matrix factorization framework.

Main Results:

• The proposed SoInp method was evaluated through comprehensive experiments on real-world datasets.
• SoInp demonstrated notable improvements in rating prediction accuracy compared to state-of-the-art models.
• The study confirmed the significant impact of implicit user influence on recommendation performance.

Conclusions:

• Implicit user influence propagation is a crucial factor in social recommendation.
• The SoInp method effectively models and leverages implicit influence for enhanced rating prediction.
• Integrating implicit influence with explicit trust information offers a promising direction for future social recommender systems.