Scaling Up Matrix Factorization with Cloud Computing for Collaborative Recommendation

Abstract

Collaborative filtering (CF) is one of the most popular and efficient recommendation methods, and matrix factorization is considered a useful technique to implement CF-based systems. To scale up the matrix factorization method for large datasets, many parallel computing techniques have been proposed. In this study, we present a new approach with different data distribution schemes to fully exploit the computing power and the memory capacity of the cloud computing platform. In addition, we perform several sets of experiments to evaluate the developed approach in a cloud computing environment, and the results show its feasibility and effectiveness.