A Recommender System for Ancient Books, Pamphlets and Paintings in Ritsumeikan Art Research Center Database

Abstract

In this paper, we develop a recommender system based on the Ritsumeikan Art Research Center (ARC) database, that stores a large number of the digital versions of ancient Japanese books, pamphlets, and paintings. We provide an easier way for novices to get access to this digital archive that was mostly used by experts before and for foreigners to know about Japanese arts and cultures. This recommender system is built as a web page. Users can select items of interest and then be provided with the most interesting items predicted by the recommender system. This system uses the restricted Boltzmann machine (RBM) for collaborative filtering to predict the most interesting items.