PMD: An Optimal Transportation-based User Distance for Recommender Systems

TL;DR

This paper introduces Preference Mover's Distance (PMD), a novel user similarity measure based on optimal transportation, which effectively utilizes all user ratings and improves recommendation accuracy in sparse data scenarios.

Contribution

The paper proposes PMD, a new transportation-based user distance measure that captures user similarity even with no co-rated items, enhancing collaborative filtering performance.

Findings

PMD outperforms state-of-the-art methods in recommendation accuracy.

PMD is effective with very sparse training data.

PMD leverages efficient Earth Mover's Distance solvers.

Abstract

Collaborative filtering, a widely-used recommendation technique, predicts a user's preference by aggregating the ratings from similar users. As a result, these measures cannot fully utilize the rating information and are not suitable for real world sparse data. To solve these issues, we propose a novel user distance measure named Preference Mover's Distance (PMD) which makes full use of all ratings made by each user. Our proposed PMD can properly measure the distance between a pair of users even if they have no co-rated items. We show that this measure can be cast as an instance of the Earth Mover's Distance, a well-studied transportation problem for which several highly efficient solvers have been developed. Experimental results show that PMD can help achieve superior recommendation accuracy than state-of-the-art methods, especially when training data is very sparse.