摘要
针对教育领域特定应用场景,利用数据挖掘技术处理教育数据是目前热点研究问题之一。课程成绩预测指对一门课程学生的期末成绩进行预测,其关键问题是通过选取合适的学生特征和确定最优的预测算法来构建预测准确率高的模型。针对大学基础课的特点,从主客观两方面选择特征,对比了4个效果最优的课程成绩预测分类算法,以准确率较高的算法构成加权投票集成算法,发现加权投票集成算法的预测准确率和AP值最高,为利用数据挖掘技术实现课程成绩预测提供了一种有效的方法。
To process the educational data from specific application scenario in education field by data mining technology is one of the popular research projects. Course performance prediction refers to the prediction of final performance of students in a curriculum,whose key issue is to build a high accuracy model by selecting proper student features and determining optimal prediction algorithm. The features are selected from both subjective and objective aspects according to features of university basic courses. Four optimal classification algorithms of curriculum performance prediction are compared and the weighted voting integrated algorithm is constituted with the algorithms with higher accuracy. The weighted voting integrated algorithm can reach the best accuracy and AP value,which provides an effective method for realization of curriculum performance prediction by utilizing data mining technology.
作者
陈佳明
骆力明
宋洁
CHEN Jiaming;LUO Liming;SONG Jie(College of Information Engineering,Capital Normal University,Beijing 100048,China)
出处
《现代电子技术》
北大核心
2020年第1期93-98,共6页
Modern Electronics Technique
基金
国家自然科学基金面上项目：面向关联数据的信息检索关键技术研究（61672361）
北京成像技术高精尖创新中心资助项目（BAICIT-2016004）
关键词
成绩预测模型
教育数据挖掘
加权投票集成算法
模型构建
大学基础课
分类算法
performance prediction model
educational data mining
weighted voting integrated algorithm
model construction
college elementary course
classification algorithm
