基于极限学习机算法的学困生预测研究  

Research on the Learning Difficulty Students Forecasting based on Extreme Learning Machine

作　　者：李刚生[1] 高铁刚[2] 刘旭[3] 于海波[4] LI Gang-sheng1 GAO Tie-gang2, LIU Xu3 ,YU Hai-bo4(1. Department of Education, Ocean UniversitY of China, Qingdao, Shandong, China 266100; 2. School of Education Technology, Shenyang Normal University, Shenyang, Liaoning, China 110034; 3. College of Engineering, Ocean University of China, Qingdao, Shandong, China 266100; 4. Basic Computer Teaching Department, Ocean University of China, Qingdao, Shandong, China 26610)

机构地区：[1]中国海洋大学教育系,山东青岛266100 [2]沈阳师范大学教育技术学院,辽宁沈阳110034 [3]中国海洋大学工程学院,山东青岛266100 [4]中国海洋大学计算机基础部,山东青岛266100

出　　处：《现代教育技术》2018年第4期34-40,共7页Modern Educational Technology

基　　金：中国海洋大学学习支持中心项目"基于极限学习机的学困生预测研究"(项目编号:2016ZZ09);基本科研业务费校文科培育类专项项目"学生学业发展增值评价研究"(项目编号:201415003)的阶段性研究成果

摘　　要：高校学困生预测方法的研究正越来越受到研究者的关注,但目前还没有一种成熟有效的学困生预测方法。针对该问题,文章提出了一种大数据环境下基于极限学习机的学困生预测方法,并以中国海洋大学2011级学生的学籍信息、心理测试得分、第一学期考试成绩为输入变量,以学生的学困情况为输出变量,进行了极限学习机的训练;同时,以2012级学生数据作为测试集输入极限学习机进行测试。测试结果表明,约有46%的学困生被准确预测,7%的非学困生被误判,此预测结果验证了文章所采用方法的有效性。Research on forecasting methods of university learning difficulty students has widely attracted researchers＇ attention, but there hasn＇t been an effective method giving convicing results at present. In this paper, a new method of forecasting learning difficulty students was proposed based on the extreme learning machine（ELM）. Studnets of 2011 in the author＇s university were selected to use their status information, psychological test scores and the first semester scores as the input variables, whereas those of the learning difficulty students were selected as the output variables and they receive the extreme learning training. Meanwhile, the data from students of 2012 are loaded into ELM as testing data. The test results show that about 46% learning difficulty students are predicted accurately, while 7% of which are misjudged. Therefore, the validity of the proposed method is verified.

关 键 词：学困生 预测 极限学习机 

分 类 号：G40-057[文化科学—教育学原理]