Automatic Estimation and Feature Word Analysis of Universities Using University Medium-term Plans

Abstract

Japanese national universities are required to draw up medium-term plans based on logic models. However, universities continue to struggle due to not knowing how to describe the five viewpoints (purpose, input, activity, output and outcome) of the logic model. In this paper, machine learning was applied to the text of third-phase medium-term plans, categorized according to the logic model, and an investigation was made into its performance in identifying each university. In addition, the feature words of the universities were analyzed, revealing the issues facing each of them.