The Study of Institutional Research of Applying Data Mining to National University Enrollment Management

碩士 === 國立暨南國際大學 === 教育政策與行政學系 === 105 === Internationalization, marketization, low birth rate and regionalization have made huge impact to Taiwan higher education. The university management is gradually marketed which is different from traditional way depending on government allowance. Institutional...

Full description

Bibliographic Details
Main Authors: CHEN, TING-YI, 陳庭逸
Other Authors: LIN, SUNG-PO
Format: Others
Language:zh-TW
Published: 2017
Online Access:http://ndltd.ncl.edu.tw/handle/sdp4c2
Description
Summary:碩士 === 國立暨南國際大學 === 教育政策與行政學系 === 105 === Internationalization, marketization, low birth rate and regionalization have made huge impact to Taiwan higher education. The university management is gradually marketed which is different from traditional way depending on government allowance. Institutional research is important for universities managers to make decisions by widely and systematic collecting and analyzing finance, management, research results, students learning results and graduates data. The purposes of this study are as follows: 1. to understand how to improve Taiwan higher education in the changing environment; 2. to explore the institutional research model of university enrollment management; 3. to present the recommendations to university enrollment management according to the research results. At first, this study explores the impact of internationalization, marketization, low birth rate and regionalization; secondly, explaining the development of institutional research and context; finally, explaining higher education enrollment management model. Be confronted with internationalization, marketization, low birth rate and regionalization, institutional research is used to be one of solving way for financial difficulties. This study is focus on enrollment management of institutional research, to be used for improving the university business strategy. This study divides enrollment management to three parts of student source analysis, learning assessment and graduation performance. These three parts is a continued and backtracking interaction process. The analysis case of this study is National Chi Nan University, analyzing its institutional research data by data mining. The data mining divided to four steps: problem definition, data aggregation, information analysis, visualization. The main analysis models are clustering and classification. Statistical analysis method of clustering is cluster analysis, statistical analysis method of classification is decision tree, and using SPSS, R and QlikView to analyze. This study used student source analysis, learning assessment and graduation performance that three steps to build enrollment management model. By clustering to know this case university’s advantage in senior high school and vocational high school. The main first year students of graduate institute are from this case university’s own university graduates. By using decision tree for learning assessment, it can build Early Warning Systems to exhume that the total passing rate and the college they attend is the key whether students are drop out. Finally, executing graduation performance by analyzing the case university and competitor employment ratio, the result is the case university employment ratio is high level, but the trend of bachelor, master and doctor employment ratio is negative growth. According to the result, this study raise three conclusions: 1. internationalization, marketization, low birth rate and regionalization make huge impact to Taiwan higher education, new innovative thinking and practice are in response to the challenges of university enrollment management. 2. institutional research promote operate performance, the most effective enrollment management is in response to internationalization, marketization, low birth rate and regionalization. 3. enrollment management is focus on the three parts: Source analysis, learning assessment and graduation performance. At last, basing on the conclusions, raising suggestions for the education authorities, university decision makers and future research.