An improved HeatS+ProbS hybrid recommendation algorithm based on heterogeneous initial resource configurations

TL;DR

This paper introduces a heterogeneity-based modification to the HeatS+ProbS hybrid recommendation algorithm, significantly enhancing its accuracy and diversity in user-object link predictions.

Contribution

It proposes a novel approach using heterogeneous initial resource configurations, improving upon the existing homogeneous setup in recommendation algorithms.

Findings

Heterogeneous initial configurations improve accuracy and diversity.

The proposed method outperforms traditional HeatS+ProbS algorithms.

Improvements are robust across different datasets.

Abstract

Network-based recommendation algorithms for user-object link predictions have achieved significant developments in recent years. For bipartite graphs, the reallocation of resource in such algorithms is analogous to heat spreading (HeatS) or probability spreading (ProbS) processes. The best algorithm to date is a hybrid of the HeatS and ProbS techniques with homogenous initial resource configurations, which fulfills simultaneously high accuracy and large diversity. We investigate the effect of heterogeneity in initial configurations on the HeatS+ProbS hybrid algorithm and find that both recommendation accuracy and diversity can be further improved in this new setting. Numerical experiments show that the improvement is robust.