Categorizing Bibliographic Data for Detection of Transition in Academic Subjects

Decision making for education at universities calls for objective analysis using reliable data. In addition to commonly used methodology, it is inspirational to introduce innovative approaches to large educational data analysis. In this study, we apply text mining techniques to analysis of a bibliography database for detecting temporal differences in academic subjects at higher education. Our approach is designed on the assumption that if automatic text categorization is effective in detecting topical distinction in book titles in different subject areas, it should also be effective to detect time-oriented distinction in the same subject area, wherein state-of-the-art books are sufficiently different from classical books. The experimental results in this study suggest that our proposed approach is feasible to detect and visualize transition in book titles of university libraries in Japan.