Information Filtering via Self-Consistent Refinement

TL;DR

This paper introduces a self-consistent refinement framework for recommender systems, improving performance and convergence speed by integrating similarity-based and spectrum-based algorithms, as validated on benchmark data.

Contribution

It presents a novel general framework that enhances recommendation accuracy and efficiency by embedding two key algorithms, a significant advance over existing methods.

Findings

Faster convergence compared to standard methods

Improved recommendation performance on benchmark data

Effective integration of similarity and spectrum-based algorithms

Abstract

Recommender systems are significant to help people deal with the world of information explosion and overload. In this Letter, we develop a general framework named self-consistent refinement and implement it be embedding two representative recommendation algorithms: similarity-based and spectrum-based methods. Numerical simulations on a benchmark data set demonstrate that the present method converges fast and can provide quite better performance than the standard methods.