Abstract To show the effectiveness or limitation of using library loan records for book recommendation, we implemented the collaborative filtering system (henceforth LLR system) which is similar to that of Harada & Masuda (2010)
In this paper, we propose a method to recommend Japanese books to university students through machine learning modules based on several features, including library loan records. We determine the most effective method among the ones that used (a) a support vector machine (SVM), (b) a random forest, and (c) Adaboost. Furthermore, we assess the most effective combination of relevant features among (1) the association rules derived from library loan records, (2) book titles, (3) Nippon Decimal Classification (NDC) categories, (4) publication years, and (5) frequencies with which books were borrowed. We conducted an experiment involving 60 subjects who were students at T University. The books recommended by our candidate methods as well as the loan records used were obtained from the T University library. The results showed that books recommended by the method that employs an SVM based on features (2), (3), and (5) were rated most favorably by subjects. The method outperforms previous book recommendation techniques, such as that proposed by Tsuji et al. (2013), and is comparable in recommendation performance to the website Amazon.co.jp. The results obtained were independent of student grades, NDC categories, and the publication years of books.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.