有鑑於國內少子化的影響，大專校院學生數劇減在民國 95年至105年間呈現負成長，招生日趨競爭，若能減少在校學生的流失，對學校而言則是一大助力。本研究以一大學 102學年度入學大學日間部資管系學生歷史學籍資料，運用資料採礦技術決策樹分析，找出多元入學學生流失率與學業表現等因素，提供相關之建議以降低學生流失。 本研究共取得有效資料 76筆，在資料特性分析發現學生學業表現以繁星推薦學生學業表現仍為各入學管道中，學業表現較為突出部分；反觀轉學考進入學校之學生學業表現仍為最弱；所以學業表現與學生入學方式有關。在學生流失率部份，與居住地區和入學方式有關。在資料採礦部分發現，以學業總成績為最重要的因素，入學方式與居住地區為其次因素，再其次為性別與班級。 In view of the impact of the domestic minority, college students in D.C. 2006 to 2016 years showed negative growth, enrollment increasingly competition, if the loss of students in school, the school is a big help. This study uses the data mining technology decision tree analysis to find out the factors such as the loss rate of the students and the academic performance of the students, and provide relevant suggestions to reduce the loss of students. In this study, a total of 76 valid data were obtained. In the analysis of the data, it was found that the academic performance of the students was still recommended by the stars. The academic performance of the students was still the most prominent part of the students. The academic performance of the students was still the weakest. So academic performance and student enrollment. In the part of the student's wastage, related to the area of residence and admission. In the data mining part of the discovery, to academic total score for the most important factor, school enrollment and residential areas as the second factor, followed by gender and class.
