SGD-Rec: A Matrix Decomposition Based Model for Personalized Movie Recommendation

Abstract

A personalized recommendation has been an active area of research. Many companies such as Facebook, Amazon, and eBay have incorporated such functionality to enhance user experience and engagement. In today’s market, streaming digital contents (e.g., online movies) have become ubiquitous and accessi-ble from anywhere and anytime. The rapid growth of streaming market urges many providers to offer a personalized experience to capture customer loyalty. In this paper, we present a movie recommending system based on our proposed rating prediction algorithm using singular value decomposition (SVD). Empirical evaluation is conducted on two tasks: rating prediction and movie recommendation, using two case studies from MovieLens and Thaiware Movie.