基于机器学习的“软助”证书挂科生分类预测研究  

Research on the Classification and Prediction of the Students Who Received the Certificate of Soft Assistance Based on Machine Learning

作　　者：何雪锋[1] HE Xue-feng(Sichuan Vocational College of Information Technology,Guangyuan Sichuan 628017,China)

机构地区：[1]四川信息职业技术学院软件学院,四川广元628017

出　　处：《河北软件职业技术学院学报》2021年第4期6-10,共5页Journal of Hebei Software Institute

基　　金：四川信息职业技术学院平台项目“成绩预警模型研究与实践”(2020KC24)。

摘　　要：针对某高校软件专业部分学生无法一次性通过软件助理工程师考证的现状,采集大一上下学期22门课程的所有成绩,结合离差标准化、SMOTE+Tomek Links过采样以及优化版的GB(Gradient Boosting)算法XGBoost(Extreme Gradient Boosting)等,构建了“软助”挂科生分类预测模型。该模型在“软助”挂科生预测中准确率较高。实验证明,采用XGBoost算法构建的模型比其他算法构建的模型效果更好,对预测可能挂科的学生提前预警,保证了软助证书的通过率。In view of the fact that some students of software major in a university cannot pass the soft assistant test at one time,by using all the results of 22 courses in the first and second semesters of the freshman year,combined with deviation standardization,SMOTE+Tomek Links oversampling,and optimized GB(Gradient Boosting)algorithm XGBoost(Extreme Gradient Boosting)etc.,the classification prediction model of soft assistant students is constructed.This method has a higher accuracy rate in the prediction of soft-assisted students.Experiments have proved that the model constructed by the XGBoost algorithm is better than the model constructed by other algorithms.It provides early warning for students who may fail to pass the course and ensures the pass rate of the certificate.

关 键 词：成绩预测 离差标准化 XGBoost 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]