Korean J General Edu Search


Korean Journal of General Education 2012;6(3):743-767.
Published online September 30, 2012.
A Study on Factors of Education’s Outcome in Liberal Computer Education Using Regression and Data Mining Analysis
Wan-Seop Kim
로지스틱 회귀분석과 데이터마이닝 분석을 이용한 컴퓨터 교양교육 성과의 요인에 대한 연구
In order to improve the outcomes of liberal arts education, a cyclic process which analyzes the factors affecting education and applies the results of the analysis to education is required. So it is very helpful to collect various data which are relevant to the educational outcome such as classroom environment, professors, student’s colleges, department, sex and analyze these data for finding factors and relations of them. In this study, we performed the work to find out the factors of computer liberal education by analyzing the data obtained from ‘Computer-Practice1’which is an essential course at Soongsil university. In this ‘Computer-Practice1’ course, MOS(Microsoft Office Specialist) certification examination is used for testing each student’s academic achievement. In the experiment of this study, the data were used, which had been obtained from the 2,865 students who took the ‘Computer-Practice1’ course in 2011. And as the analysis method, logistic regression and decision tree analysis were used. We could find out the various factors which affected the outcome of the education by applying regression analysis which is the traditional statistical method. But because the hierarchical relationship among the factors could not be investigated by using regression analysis, we used the decision tree analysis which is the representative method in data-mining. There is a big advantage to use decision tree analysis because this method provides a tree model which explains the relationship of the factors that affect the dependent variable. And this tree model could be interpreted intuitively and provide useful applicable rules. In particular, three algorithms including QUEST, CRT and CHAID are used in order to overcome the disadvantage which is weak in correlation of the input variables through decision tree analysis. Through the experiment, I found out that various factors including education period, professors, classrooms and class hours affect the educational outcome complexly.
Key Words: Liberal Computer Education, Factors of Education’s outcome, Logistic Regression, Data-mining, Decision tree


Browse all articles >

Editorial Office
203-827. Chung-Ang University
84, Heukseok-ro, Dongjak-gu, Seoul, Republic of Korea, 06974
Tel: +82-2-820-5949    E-mail: hangang331@cau.ac.kr                

Copyright © 2022 by The Korean Association of General Education.

Developed in M2PI

Close layer
prev next