基于K-means融合决策树分类算法的学生表现评价模型构建  

Construction of student performance evaluation model based on K-means fusion decision tree classification algorithm

作　　者：许亚杰 梁靖涵 Xu Yajie;Liang Jinghan(College of Information Engineering,Zhengzhou University of Science&Technology,Zhengzhou 450000,China)

机构地区：[1]郑州科技学院信息工程学院,河南郑州450000

出　　处：《无线互联科技》2022年第22期134-137,共4页Wireless Internet Technology

摘　　要：针对当前初级中学教育过于注重学生智力因素的研究,而忽略非智力因素影响的问题,文章从影响学生表现的家庭、个人和学校特征出发,提出了一种基于K-means融合决策树分类算法的学生表现评价模型。该模型首先通过Kmeans算法将学生表现划分成warning和keeping两类;在此基础上,设定树深度为3,最小拆分节点样本数为25,建立CART模型。从决策树模型可以看出影响学生表现的重要指标有外出次数、健康状况和旷课次数。实验结果表明:学生如果外出次数过多,且身体健康状况较好,经常旷课,会严重影响学生表现。因此,在学生教育中,家长要合理地限制学生外出次数,学生要保持身体健康,学校要严格落实考勤,以此保障学生能够投入学习,提高学习效果。Aiming at the problem that the current junior high school education pays too much attention to the research of students’ intelligence factors, while ignoring the influence of non-intelligence factors, this paper starts with the family,personal and school characteristics that influence student performance, proposes a student performance evaluation model based on K-means fusion decision tree classification algorithm. The model first divides students’ performance into two categories: warning and keeping through K-means algorithm;on this basis, the tree depth is set to 3, and the minimum number of split node samples is 25, and the CART model is established. From the decision tree model, it can be seen that the important indicators that affect the performance of students are the number of going out, health status and the number of absenteeism. The experimental results show that if students go out too many times, and their physical health is good, they often miss classes, which will seriously affect their performance. Therefore, in student education, parents should reasonably limit the number of times students go out, students should maintain good health, and schools should strictly implement attendance, so as to ensure that students can devote themselves to learning and improve learning effects.

关 键 词：学生表现评价 K-MEANS 决策树 CART算法 

分 类 号：G632.0[文化科学—教育学] TP311.13[自动化与计算机技术—计算机软件与理论]