A Style-Aware Collaborative Filtering-Based Recommender System

Abstract

Online shopping for clothing products is growing rapidly. In order to avoid choice overload and match consumers with the most suitable products, retailers use recommender systems. However, unlike other products, recommending clothes can be challenging. Most customers not only search a clothes by their popularity or price but also by style. We present a Collaborative Filtering recommender system based on the traditional Matrix Factorization which incorporates items’ contextual information in order to discover users’ aesthetic preferences. We apply a style-aware recommender model in a real-world dataset of Amazon for experimental evaluation, demonstrating that our algorithm outperforms the state-of-the-art CF-based recommender approach.

Keywords: Recommender Systems, E-commerce, Collaborative Filtering