Collaborative Filtering Approach to Link Prediction

TL;DR

This paper introduces a collaborative filtering-based enhancement framework for local similarity indices in link prediction, significantly improving their accuracy and robustness across different networks with lower complexity.

Contribution

It proposes a novel SCF framework that enhances local indices for link prediction, offering a more robust and computationally efficient alternative to complex algorithms.

Findings

SCF significantly improves local indices' accuracy.

Enhanced indices are more robust across various networks.

The combined approach achieves competitive performance with lower complexity.

Abstract

Link prediction is a fundamental challenge in network science. Among various methods, local similarity indices are widely used for their high cost-performance. However, the performance is less robust: for some networks local indices are highly competitive to state-of-the-art algorithms while for some other networks they are very poor. Inspired by techniques developed for recommender systems, we propose an enhancement framework for local indices based on collaborative filtering (CF). Considering the delicate but important difference between personalized recommendation and link prediction, we further propose an improved framework named as self-included collaborative filtering (SCF). The SCF framework significantly improved the accuracy and robustness of well-known local indices. The combination of SCF framework and a simple local index can produce an index with competitive performance and much lower complexity compared with elaborately-designed state-of-the-art algorithms.