A Study on the Applications of Data Mining Techniques to Prevent Student Dropouts – Based on a TVE Further Education College

• Main Authors: Ruei-Lin Jhao, 趙瑞麟
• Other Authors: Huan-Ming Chuang
• Format: Others
• Language: zh-TW
• Published: 2007
• Online Access: http://ndltd.ncl.edu.tw/handle/90907092394322122489

碩士 === 國立雲林科技大學 === 資訊管理系碩士班 === 95 === In order to comply with the national promoting policy of lifetime learning, most TVE institutes in Taiwan have made full use of their educational resources to offer adult further education programs. Since courses of these programs are scheduled in night time or weekends mostly, employed workers now have the opportunity to balance their job and further studies at the same time. Nevertheless, since all institutes are expanding their recruitment scale, they need to face the challenges of short demand. As a result, how to align further education with needs of both students and industries becomes an important issue. This research bases on a private institute of technology located in central Taiwan, with data collected from 2001 to 2005 student enrollments. It aims to reach the following objectives: (1) investigating essential reasons of student dropouts, (2) establishing effective model for predicting student dropouts, and (3) suggesting strategies for retarding student dropouts. Data mining techniques as C5.0 and CART decision tree algorithms and neural network are conducted by SPSS Clementine 7.2. Major findings can be listed as follows. First, C5.0 decision tree algorithm performs better in predicting student dropouts. Second, major reasons for student dropouts are without semester register, without resuming interrupted studies, and active dropout due to personal job factor. Third, according to decision tree algorithms, major factors to categorize dropouts are conduct grades, age, and academic grades in turn generally. For those predicated as with high propensity to dropout, follow-ups and counseling can be conducted to prevent them from actual dropouts. Besides, these findings can also provide helpful guidance for the further education college to formulate appropriate strategies regarding recruitment, adjusting offered education programs, and arranging courses in programs.