Coupled Item-based Matrix Factorization

TL;DR

This paper introduces CIMF, a matrix factorization method that incorporates coupled item attributes to better handle cold start and sparsity issues in recommender systems.

Contribution

It proposes a novel coupled similarity measure for item attributes and integrates it into matrix factorization, addressing limitations of iid attribute assumptions.

Findings

CIMF outperforms benchmark methods on open datasets.

The coupled similarity measure captures implicit attribute relationships.

The approach improves recommendation accuracy in cold start scenarios.

Abstract

The essence of the challenges cold start and sparsity in Recommender Systems (RS) is that the extant techniques, such as Collaborative Filtering (CF) and Matrix Factorization (MF), mainly rely on the user-item rating matrix, which sometimes is not informative enough for predicting recommendations. To solve these challenges, the objective item attributes are incorporated as complementary information. However, most of the existing methods for inferring the relationships between items assume that the attributes are "independently and identically distributed (iid)", which does not always hold in reality. In fact, the attributes are more or less coupled with each other by some implicit relationships. Therefore, in this pa-per we propose an attribute-based coupled similarity measure to capture the implicit relationships between items. We then integrate the implicit item coupling into MF to form the Coupled Item-based Matrix Factorization (CIMF) model. Experimental results on two open data sets demonstrate that CIMF outperforms the benchmark methods.