Collaborative filtering with diffusion-based similarity on tripartite graphs

TL;DR

This paper introduces a diffusion-based similarity measure for collaborative filtering that leverages user, object, and tag relationships in tripartite graphs, improving recommendation accuracy.

Contribution

It proposes a novel diffusion-based similarity measure on tripartite graphs, enhancing personalized recommendations by integrating preference and tagging information.

Findings

Outperforms cosine similarity in ranking score

Achieves higher Recall and Precision

Effective in personalized recommendation tasks

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.