Improved Collaborative Filtering Algorithm via Information Transformation

Abstract

In this paper, we propose a spreading activation approach for collaborative filtering (SA-CF). By using the opinion spreading process, the similarity between any users can be obtained. The algorithm has remarkably higher accuracy than the standard collaborative filtering (CF) using Pearson correlation. Furthermore, we introduce a free parameter $\beta$ to regulate the contributions of objects to user-user correlations. The numerical results indicate that decreasing the influence of popular objects can further improve the algorithmic accuracy and personality. We argue that a better algorithm should simultaneously require less computation and generate higher accuracy. Accordingly, we further propose an algorithm involving only the top-$N$ similar neighbors for each target user, which has both less computational complexity and higher algorithmic accuracy.

5 pages, 4 figures

References in corpus (6)

• How to project a bipartite network?
• Heat Conduction Process on Community Networks as a Recommendation Model
• Power-law Strength-Degree Correlation From a Resource-Allocation Dynamics on Weighted Networks
• Effect of initial configuration on network-based recommendation
• Spreading gossip in social networks
• Recommendation model based on opinion diffusion

Cited by in corpus (11)

• Recommender Systems
• Personalized Recommendation via Integrated Diffusion on User-Item-Tag Tripartite Graphs
• Information filtering via biased heat conduction
• Network-based recommendation algorithms: A review
• Information filtering based on transferring similarity
• Highly accurate recommendation algorithm based on high-order similarities
• Solving the accuracy-diversity dilemma via directed random walks
• Effect of user tastes on personalized recommendation
• Degree correlation effect of bipartite network on personalized recommendation
• Collaborative filtering based on multi-channel diffusion
• Ultra accurate collaborative information filtering via directed user similarity