摘要
大学生的情绪问题已经成为高校学生管理的难点,如何让学生管理工作部门尽早发现学生情绪变化并提前介入是当前急需解决的问题.针对大学生情绪变化检测问题,利用手机、可穿戴设备等收集了学生运动、环境、状态等数据,结合情感模型对学生情感进行了度量.提出了基于因子分解机的深度交叉网络(FDCN)情绪变化检测模型,该模型以深度交叉网络为基础,结合因子分解机充分学习特征的低阶组合和高阶组合及非线性关系,提高了对大学生情绪变化检测的精度.采用真实数据集进行实验的结果说明:该模型对比其它模型在多个评价指标上均有提升,具有更准确和更稳定的预测效果.
The emotional problems of college students have become a difficult problem of college student management,and how to enable student management departments to detect the emotional changes of students as soon as possible and intervene in advance is a current urgent problem.The problem of detecting emotional changes of college students is addressed by using cell phones and wearable devices to collect data of students′movement,environment,status and other data,combine with the emotion model to measure students′emotions,a deep crossover network(FDCN)emotional change detection model based on factorization machine is proposed,which is based on deep crossover network,combined with factorization machine to fully learn the low-order combination and high-order combination of features and nonlinear relationships,so as to improve the accuracy of college students′emotional change detection.The results of experiments using real data sets illustrate that the model has improved in several evaluation indexes compared with other models,and has more accurate and stable prediction effects.
作者
陈宇
邵军
赖文天
CHEN Yu;SHAO Jun;LAI Wentian(School of Computer&Hubei Education Cloud Service Engineering Technology Research Center,Hubei University of Education,Wuhan 430205,China;Hubei Water Resources Technical College,Wuhan 430071,China)
出处
《中南民族大学学报（自然科学版）》
CAS
北大核心
2022年第4期483-489,共7页
Journal of South-Central University for Nationalities：Natural Science Edition
基金
湖北省教育厅科学技术研究计划中青年人才资助项目(Q20143005)。
