A Mobile Application for Personalized Movie Recommendations with Dynamic Updates
Harris Papadakis, Paraskevi Fragopoulou, Nikos Michalakis, Costas Panagiotakis · 2018
Recommender systems try to provide users with accurate personalized suggestions for items based on an analysis and assessment of previous user behaviour and satisfaction. With the abundance of online products and services, the development of methods and tools that successfully identify users preference patterns is now, more than ever, imperative. In this paper we present MovieScore, an online mobile application for personalized movie recommendation, which predicts movies of interest for users based on previously provided user satisfaction ratings. The back-end of the application relies on a novel recommendation algorithm that has been shown to provide excellent results and exhibits dynamic incremental update behavior. This easy-to-use application allows users to effortlessly specify their preferences by rating already watched movies. In turn, the application employs the aforementioned state-of-the-art algorithm in order to provide users with accurate, personalized movie recommendations. The proposed system exhibits dynamic update behavior, characteristic that renders it ideal for use in mobile devices. When new items arrive (movies, users, or ratings) the algorithm is retrained incrementally which has negligible impact on its performance. We describe the design, implementation details and functionality of the mobile application as well as the basics of the underlying recommendation algorithm.