基于校园一卡通与学生基本信息预测落后生 Predicting At-Risk Students Based on the Campus Card and Students’ Basic Information

准确预测大一新生中的落后生对提高高校毕业率有重要的影响。本文研究了校园一卡通打卡记录和学生成绩之间的关系，并结合学生基本信息和前期成绩记录，提出一种基于机器学习与统计学的预测落后生的方法。在中国海洋大学2013级3680名新生共计419.4万条记录上进行了实验，结果显示所提出的方法查全率达到52%，查准率达到77%，具有较好的实用性。 Accurate prediction of at-risk students in freshmen is extremely important to improve the graduation rate of a university. This paper explores the relationships between the campus card records and the performance of students. In addition, combining with the basic information of students and their previous exam scores, the paper proposes a method of predicting at-risk students based on machine learning and statistics. The experiment conducts on a dataset with 4.194 million items of 3680 freshmen of grade 2013 from the Ocean University of China, and the result shows that the recall rate is 52% and the precision is 77%, with good practical performance.