Improved Collaborative Filtering Algorithm via Information Transformation

TL;DR

This paper introduces an improved collaborative filtering algorithm using information transformation and spreading activation, achieving higher accuracy and efficiency by regulating object influence and focusing on top-N neighbors.

Contribution

The paper presents a novel spreading activation-based collaborative filtering method with a tunable parameter and a top-N neighbor approach for better accuracy and reduced computation.

Findings

Higher accuracy than standard CF using Pearson correlation

Decreasing influence of popular objects improves personalization

Top-N neighbor approach reduces computation while increasing accuracy

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.