数据挖掘算法在大学生成绩分析中的应用  

Application of data mining algorithm in college student achievement analysis

作　　者：何韵竹 姚登举[1] 张敬辉 李哲[1] 

机构地区：[1]哈尔滨理工大学软件学院,黑龙江哈尔滨150040

出　　处：《高师理科学刊》2017年第12期30-34,共5页Journal of Science of Teachers＇College and University

基　　金：黑龙江省大学生创新创业训练计划项目(201610214058)

摘　　要：数据挖掘技术可以从大量的数据中通过算法发现隐藏其中的未知信息,被广泛应用到各种数据分析任务中,然而教育数据挖掘领域的研究目前仍然很少.针对高校大学生成绩分析问题,研究了经典的数据挖掘算法的分类性能,包括支持向量机、多元自适应回归样条、K-最近邻、决策树和随机森林.基于随机森林算法的变量重要性度量,分析了软件工程专业特定的课程对于学生综合学习效果的影响和作用.在真实的大学生学习数据集上进行了实验,获得了一些有意义的发现.The technology of data mining, which can discover unknown information hidden in a large amount of data by algorithm, is widely applied to various data analysis tasks. However, data mining research in the field of education is still rare. Aiming at the achievement analysis of college students, the classification performance of classical data mining algorithms is studied, including support vector machines, multivariate adaptive regression splines, K- nearest neighbors, decision trees and random forests. Based on the variable importance measurement of random forest algorithm, analyzes the influence and function of specific courses of software engineering specialty on students＇ comprehensive learning effect. Experiments were carried out on real university learning data sets, and some significant results were obtained.

关 键 词：学习成绩分析 数据挖掘 随机森林 支持向量机 K-最近邻 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]