摘要
在我国社会、经济与科学技术不断发展的背景下,信用成为生活中不可或缺的重要元素,但学术造假、信贷违约等大学生信用问题也逐渐显露出来。本文对我国大学生学术信用及相关内容的研究现状进行了梳理,总结了现阶段其存在的问题。基于合成少数类过采样技术(SMOTE)与线性逻辑回归(LR),设计了SMOTE-LR学术失信预测模型,通过实证分析,该预测模型结果具有一定的准确性。大学生学术失信预测模型在人才选拔、学历考试及个人征信等场景应用中起着重要的作用。
In the context of the continuous development of economy and technology,credit has become an indispensable important element in daily life.But credit issues,such as academic fraud and credit defaults,have gradually emerged.This article sorts out the current research status of academic credit and related content of college students in China,and summarizes the existing problems at this stage.Then,based on the Synthetic Minority Oversampling Technique(SMOTE)and Linear Logistic Regression(LR),the SMOTE-LR academic default prediction model was designed.Through empirical analysis,the results of the prediction model have certain accuracy.The predictive model of college students academic default will play an important role in multi-scenario applications such as talent selection in social activities,academic qualification examinations,and personal credit investigation.
作者
李志强
余炫朴
LI Zhi-qiang;YU Xuan-pu(College of Statistics,Jiangxi University of Finance and Economics,Nanchang 330013,China)
出处
《东北师大学报（哲学社会科学版）》
CSSCI
北大核心
2021年第4期76-83,共8页
Journal of Northeast Normal University(Philosophy and Social Science Edition)
基金
全国统计科学研究项目(2020LY038)
国家自然科学基金委员会资助项目(71663024)
江西省高校人文社科研究基地重点研究项目(JD1573)。
关键词
学术信用
失信预测
合成少数类过采样技术
线性逻辑回归
Academic Credit
Prediction of Default
Synthetic Minority Oversampling Technique
Linear Logistic Regression
作者简介
李志强,男,江西财经大学统计学院教授、博士生导师;余炫朴,男,江西财经大学统计学院博士研究生,南昌师范学院讲师。
