Collaborative filtering with diffusion-based similarity on tripartite graphs

Abstract

Collaborative tags are playing more and more important role for the organization of information systems. In this paper, we study a personalized recommendation model making use of the ternary relations among users, objects and tags. We propose a measure of user similarity based on his preference and tagging information. Two kinds of similarities between users are calculated by using a diffusion-based process, which are then integrated for recommendation. We test the proposed method in a standard collaborative filtering framework with three metrics: ranking score, Recall and Precision, and demonstrate that it performs better than the commonly used cosine similarity.

8 pages, 4 figures, 1 table

References in corpus (7)

• Predicting Missing Links via Local Information
• Personalized Recommendation via Integrated Diffusion on User-Item-Tag Tripartite Graphs
• Power-law Strength-Degree Correlation From a Resource-Allocation Dynamics on Weighted Networks
• Information filtering based on transferring similarity
• Personal Recommendation via Modified Collaborative Filtering
• Information Filtering via Self-Consistent Refinement
• Empirical analysis on a keyword-based semantic system

Cited by in corpus (10)

• Recommender Systems
• Tag-Aware Recommender Systems: A State-of-the-art Survey
• Information filtering via biased heat conduction
• Hypergraph model of social tagging networks
• Graph-based Collaborative Ranking
• Information Filtering via Collaborative User Clustering Modeling
• Recommending investors for new startups by integrating network diffusion and investors' domain preference
• A Fast Recommendation Algorithm for Social Tagging Systems : A Delicious Case
• Leveraging tagging and rating for recommendation: RMF meets weighted diffusion on tripartite graphs
• Effective Personalized Recommendation in Collaborative Tagging Systems