Text Mining of the Midterm Goals of National Universities for the Development of Their Individual Characteristics

Abstract

This paper attempts to discover the distinguishing characteristics of Japanese national universities, and to determine the strengths of each university by analyzing their midterm goals. We propose a method that extracts words from documents, analyzes them, and visualizes the features of documents on the basis of these extracted words. We developed two systems, a cross-tabulation search engine and a keyword map search engine, for the midterm goal document. In these systems, the user both decides the viewpoint to be compared and analyzed and inputs the search word, so that the relation of the words related to the search word is visualized, and a novel comparison analysis method among university organizations is realized. In this paper, we report the outline of these systems and the results of the verification of the two systems, using "regional contribution" as an example.

An Analysis of Chronological Changes Based on Feature Words Using the Cross-Tabulation Search Engine

“…National and public universities are required to disclose their medium-term plan documents to all universities. As a result, despite the slightly different format of the document, the result of analyzing the two-dimensional map when searching for the characteristics of national universities with the keyword "region (community)" has been confirmed to be a valid result [2][7] [8]. In addition, two-dimensional visualization of public universities has yielded useful results [9] [10].…”

“…As a previous study directly related to the present study, there is a cross-tabulation search engine for the medium-term plan, which was mentioned in "2. Related work" [2]. As shown in Figure 1, typing "z" in the box will display all search results.…”

“…In Figure . 1, it is an example of setting "v" to "view", "h" to "year", "feature" to "5". See reference [2] for details of the cross-tabulation search engine.…”

“…We have checked the feature words and conducted university analysis and inter-university comparisons. We have developed a system that utilizes text mining technology to reduce this enormous amount of labor cost [2]. We were able to greatly improve efficiency and obtain new knowledge that could not be discovered by human eyes without a massive amount of work by using the developed system "Cross-Tabulation Search Engine" [2].…”

“…We have developed a system that utilizes text mining technology to reduce this enormous amount of labor cost [2]. We were able to greatly improve efficiency and obtain new knowledge that could not be discovered by human eyes without a massive amount of work by using the developed system "Cross-Tabulation Search Engine" [2]. By utilizing the cross-tabulation search engine and applying it to university self-assessment documents, it is possible to efficiently set hypotheses about chronological changes without reading a huge number of documents.…”