摘要
在高职教育中,一个能够以较高精度预测学生考试成绩的模型可帮助教师预判教学效果,从而不断改进教学方法,提高教学质量。利用正交核最小二乘法搭建了高职学生考试成绩预测模型,以学生自身特点和平时表现等信息构成模型输入特征变量,预测学生的考试成绩是否合格以及具体分数。以深圳信息职业技术学院学生作为研究对象,将模型从预测精度和稀疏度两方面与支持向量机进行比较。实验结果表明,利用正交核最小二乘法建立的模型具有良好的泛化能力,虽然拟合精度略逊于支持向量机,但是能够取得更好的分类精度和更大的稀疏度。
In the higher vocational education,a model to predict students’exam results with high accuracy can help teachers prejudge the teaching effect,hence improving their teaching techniques and enhancing the teaching quality. This paper aims to construct a mod-el to predict the students’exam results in the higher vocational college by using the orthogonal kernel least squares algorithm (OKLSA). The students’own characteristics and studying performance are employed to form the features to be input into the model and the exam results are output from the model thereafter. The information related to students from the Shenzhen Institute of Informa-tion Technology is collected and grouped into the experimental dataset. The model’s performance is tested by comparing OKLSA with Support Vector Machines(SVMs)with respect to the prediction accuracy and the structure sparsity. The experimental results show that OKLSA has good generalization ability. Although the fitting accuracy of OKLSA is lower than that of SVM,OKLSA has higher classifi-cation accuracy and sparser structure than SVM.
作者
杨东海
吴月瑞
YANG Dong-hai;WU Yue-rui(Management School,Shenzhen Institute of Information Technology,Shenzhen 518172,China)
出处
《软件导刊》
2019年第6期143-146,150,共5页
Software Guide
基金
深圳信息职业技术学院第六批校级教研项目(2016jgyb04)
广东省一流高职院校建设项目(2017)
关键词
正交核最小二乘法
高职教育
考试成绩预测
orthogonal kernel least squares algorithm
higher vocational education
exam results prediction
