Coherence analysis of research and education using topic modeling

Research and education are organically connected in that lectures convey the results of research, which is frequently initiated by inspiring lectures. As a result, the contents of lecture materials and research publications and the research capabilities of universities should be considered in the investigations of the relationships between research and teaching. We examine the relationship between research and teaching using automatic text analysis. In particular, we scrutinize the relatedness of the content of research papers with the content of lecture materials to investigate the association between teaching and research. We adopt topic modeling for the correlation analysis of research capabilities and the reflectiveness of research topics in lecture materials. We select the field of machine learning as a case study because the field is contemporary and because data related to teaching and research are easily accessible via the Internet. The results reveal interesting characteristics of lecture materials and research publications in the field of machine learning. The research capability of an institute is independent of the lecture materials. However, for introductory courses, teaching and research measures showed a weak negative relationship, and there is little relationship between the measures for advanced courses.