周慶 王衛芳 葛亮 肖逸楓 唐代

摘要:針對大學生存在學業風險、高校管理難度增大的問題,提出了基于一卡通數據與課程分類預測學生是否存在及格風險的方法。首先對計算機學院學生的一卡通數據與課程成績進行預處理與特征提取,利用皮爾遜相關系數與Apriori算法分析不同學期課程成績間、早餐次數與成績間的相關性和關聯性。然后結合早餐次數與同類型課程的成績,運用多種分類器預測學生未來成績是否及格。結果表明,該方法可預測學生某門課程是否存在不及格風險,便于教輔人員及時干預學業困難學生。
關鍵詞:學業風險;相關性分析;關聯分析;課程分類;成績預測
中圖分類號:TP391 文獻標識碼:A 文章編號:1009-3044(2018)24-0236-04
Abstract: Aiming at the academic risk of College students and the difficulty of Teaching Management in Universities,this paper presented a method based on campus card data and the idea of curriculum classification to predict whether students can pass the course examination. First of all, preprocessing the data of students' campus card data and course performance, and extracting the features, Secondly, using Pearson's correlation coefficient and Apriori algorithm to analyze not only the correlation between the course results of different semesters, but also the relevance between breakfast time and course performance. Then, combining the number of breakfast with the results of the same type of course, a variety of classifiers were used to predict whether the students' future performance was passed. The result shows that this method can predict whether there is a risk of failure in a student's course, and it is convenient for teachers to help students with academic difficulties in time.
Key words: academic risk; correlation analysis; association analysis; curriculum classification; performance prediction
1 引言
大學生在校期間的學業表現是影響其畢業及未來就業的關鍵因素,為了完善高校學生管理制度,各大高校全面推進“學業預警”制度,最終目的是幫助存在學業風險的同學順利完成學業[1]。影響學生成績的因素較多,比如師資水平、性別差異、課程難易程度、不適應大學生活等。目前,已有較多的學者針對學業成績預測做了相關研究,如武彤等人利用決策樹算法分析學生課堂表現與性別的差異,預測學生最終是否通過某門課程[2];王凱成等人利用Microsoft SQL Server提供的數據挖掘功能分析學生歷史成績數據,預測學生的平均學分績點,幫助老師提前采取措施干預學分績點不達標的同學[3];張紅林等人分析學生早餐習慣與課程成績的關聯規則,發現早餐就餐次數多、時間早的學生成績明顯高于就餐次數少、時間晚的規律[4]。……