摘要
基于数据智能分析的学习资源推送是精准支持个性化学习的教学服务方式之一。随着人工智能技术和学习分析技术的发展,通过对行为数据、测评数据和日志数据等的分析进行资源推送虽已有较成功应用,但未能实现学习者个人情感状态为引导的资源推送。为此,该研究针对个性化学习资源推荐中情感价值、情感控制理论和实践的缺失,以实现个性化学习多维度情感识别为目标,采用BERT模型和TextCNN构建个性化学习资源推荐文本情感识别模型,并提出了基于学习者作业、论坛内容等的文本情感识别模型实现过程。学习资源推荐文本情感识别模型和实现过程可为真实应用提供方法支持和技术路线指引。
Learning resource pushing based on data intelligence analysis is one of the teaching services that precisely support personalized learning.With the development of artificial intelligence technology and learning analytics,resource pushing through the analysis of behavioral data,assessment data and log data has been more successfully applied,but it fails to realize resource pushing guided by learners’personal emotional state.Therefore,this study addresses the lack of theory and practice of emotion value and emotion control in personalized learning resource recommendation,and aims to achieve multidimensional emotion recognition for personalized learning,uses BERT model and TextCNN to build a text emotion recognition model for personalized learning resource recommendation,and proposes the process of implementing the text emotion recognition model based on learners’homework and forum content.The learning resource recommendation text sentiment recognition model and the implementation process can provide methodological support and technical route guidance for real applications.
作者
许桂芳
穆肃
Xu Guifang;Mu Su(School of Information Technology in Education,South China Normal University,Guangzhou 510631,Guangdong;Institution of AI in education,South China Normal University,Guangzhou 510631,Guangdong)
出处
《中国电化教育》
北大核心
2023年第5期105-112,共8页
China Educational Technology
基金
国家自然科学基金2022年度重点项目“课堂流媒体跨模态知识元协同解析与评估方法”(项目编号:62237001)研究成果。
关键词
学习资源推荐
文本情感识别
个性化学习
recommendation of learning resources
text emotion recognition
personalized learning
