[1]王 娜,李劲松,潘子尧,等.基于特征选择的学位预警方法研究[J].计算机技术与发展,2023,33(09):24-29.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 004]
　WANG Na,LI Jin-song,PAN Zi-yao,et al.Research on Degree Early Warning Method Based on Feature Selection[J].,2023,33(09):24-29.[doi:10. 3969 / j. issn. 1673-629X. 2023. 09. 004]

点击复制

基于特征选择的学位预警方法研究()

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

33

期数:

2023年09期

页码:

24-29

栏目:

大数据与云计算

出版日期:

2023-09-10

文章信息/Info

Title:

Research on Degree Early Warning Method Based on Feature Selection

文章编号:

1673-629X(2023)09-0024-06

作者:

王 娜¹ ; 李劲松¹ ; 潘子尧² ; 姚明海^1*

1. 渤海大学 信息科学与技术学院,辽宁 锦州 121013;

2. 渤海大学 数学科学学院,辽宁 锦州 121013

Author(s):

WANG Na1 ; LI Jin-song1 ; PAN Zi-yao2 ; YAO Ming-hai1*

1. School of Information Science and Technology,Bohai University,Jinzhou 121013,China;

2. School of Mathematical Science,Bohai University,Jinzhou 121013,China

关键词:

教育数据挖掘; 特征选择; 学位预警; 支持向量机; 成绩预测

Keywords:

education data mining; feature selection; degree early warning; support vector machines; performance prediction

分类号:

TP301

DOI:

10. 3969 / j. issn. 1673-629X. 2023. 09. 004

摘要:

高校学生能够顺利获得学位,不仅对其个人就业发展至关重要,也是衡量高校教学质量的重要指标之一。 学位预警是教育数据挖掘的重要应用之一,通过学位预警可以尽早地对学生的学业情况进行警示,学生能够及时调整学习状态和方法,同时准确的学位预警也可以为改进教学指导策略提供参考依据。 现有的预警模型构建多是基于全部成绩数据,忽略了课程间的冗余性,使得构建的模型精度不足。 因此,提出基于 Fisher 特征选择方法构建学位预警模型。 利用 Fisher得分对特征进行初步筛选;然后,利用筛选后的特征构建学位预警模型;最后,通过预警模型对获得学位情况进行预测。为检验方法的有效性,在某高校汉语言文学、化学、数学与应用数学等专业真实数据上进行了大量实验。 实验结果表明,基于特征选择的学位预警方法具有良好的准确度和实用性,可以为高校学生的学位预警工作提供数据支持。

Abstract:

The successful acquisition of a degree by college students is not only crucial to their personal employment development,but alsoone of the important indicators to measure the quality of college teaching. Degree early warning is one of the important applications ofeducational data mining. The degree warning can warn students?

of their degree information as early as possible,student can adjust theirlearning state and methods in time. At the same time,accurate degree warning can provide reference for improving teaching guidancestrategies. The existing early warning models are mostly built based on all performance data, which makes the accuracy of theconstructed model is insufficient. Therefore, the degree early warning model based on Fisher feature selection method is proposed.Firstly,Fisher’s score is used to preliminarily screen the features. Then,the degree warning model is built with the selected features.Finally,the degree obtaining situation is predicted through the early warning model. In order to test the effectiveness of the proposedmethod,a large number of experiments were carried out on the real data of seven majors of Chinese language and literature major,chemistry major,and mathematics and applied mathematics in university. The experimental results show that the proposed degree earlywarning method based on feature selection has excellent accuracy and practicality, and can provide data support for the degree earlywarning of college students.

常用功能