An Issue-oriented Syllabus Retrieval System based on Terminology-based Syllabus Structuring and Visualization

Recently, scientific knowledge has grown explosively because of rapid advancements that have occurred in academia and society. Because of this dramatic expansion of knowledge, learners and educators sometimes struggle to comprehend the overall aspects of syllabi. In addition, learners may find it difficult to discover appropriate courses of study from syllabi because of the increasing growth of interdisciplinary studies programs. We believe that an issueoriented syllabus structure might be more efficient because it provides clear directions for users. In this paper, we introduce an issue-oriented automatic syllabus retrieval system that integrates automatic term recognition as an issue extraction, and similarity calculation as terminology-based document clustering. We use automatically-recognized terms to represent each lecture in clustering and visualization. Retrieved syllabi are automatically classified based on their included terms or issues. The main goal of syllabus retrieval and classification is the development of an issue-oriented syllabus retrieval website that will present users with distilled knowledge in a concise form. In comparison with conventional systems, simple keyword-based syllabus retrieval is based on the assumption that our methods can provide users, and, in particular, novice users (students), with efficient lecture retrieval from an enormous number of syllabi. The system is currently in practical use for issue-oriented syllabus retrieval and clustering for syllabi for the University of Tokyo’s Open Course Ware and for the School/Department of Engineering. Usability evaluations based on questionnaires used to survey over 100 students revealed that our proposed system is sufficiently efficient at syllabus retrieval.