Metadata Embeddings for User and Item Cold-start Recommendations

TL;DR

This paper introduces a hybrid matrix factorisation model that leverages content features to improve cold-start and sparse data recommendations, while also capturing semantic information in feature embeddings.

Contribution

The paper presents a novel hybrid matrix factorisation approach that effectively utilizes content features for improved cold-start recommendations and semantic feature embeddings.

Findings

Outperforms collaborative and content-based models in cold-start scenarios

Performs comparably to pure collaborative models with abundant data

Feature embeddings encode semantic information useful for related tasks

Abstract

I present a hybrid matrix factorisation model representing users and items as linear combinations of their content features' latent factors. The model outperforms both collaborative and content-based models in cold-start or sparse interaction data scenarios (using both user and item metadata), and performs at least as well as a pure collaborative matrix factorisation model where interaction data is abundant. Additionally, feature embeddings produced by the model encode semantic information in a way reminiscent of word embedding approaches, making them useful for a range of related tasks such as tag recommendations.