摘要
在线教育师生情感缺失问题是当前教育研究亟待解决的难题之一。在线教育弹幕文本作为学习者对在线课程内容及自身学习状态的实时反馈,隐含了大量情感信息,对于上述问题的解决具有重要意义。然而,目前鲜有针对在线教育弹幕文本开展情感分析的研究。基于此,设计了一种融合变式情感词典与深度学习技术的在线教育弹幕情感智能识别模型。具体而言:通过构建弹幕种子情感词集,计算弹幕文本与情感种子短语间的相似度,实现极短弹幕文本情感识别;借助BERT动态表征弹幕文本,双向长短时记忆网络挖掘弹幕文本中的深层特征,实现常规弹幕文本情感识别;融合两类弹幕文本情感识别信息并更新后,完成在线教育弹幕文本情感信息的智能识别。研究通过模型对比实验检验模型性能有效性,并借助具体案例验证模型应用可行性。
The lack of teacher-student sentiment in online education is a pressing issue in current educational research.Online education barrage texts,as real-time feedback from learners on the course content and their own learning status,imply a large amount of sentiment information,which has significant implications for solving the above-mentioned problem.However,there are few studies on sentiment analysis of online education barrage texts.Based on this,designing a sentimental information intelligent recognition model which merges variant sentiment dictionary and deep learning for online education barrage.Specifically,we construct a barrage seed sentiment phrases set,calculate the similarity between barrage text and sentiment seed phrases,achieving extremely short barrage text sentiment recognition;Using BERT to dynamically characterize the barrage text,bidirectional long short-term memory neural network to mine the deep features in the barrage text,achieving regular barrage text sentiment recognition;fusing and updating two types of barrage text sentiment recognition information to complete the intelligent recognition of online education barrage text sentiment.The study verified the performance validity of the model through model comparison experiments,displayed the application feasibility of the model with the help of specific cases.
作者
李浩君
汪旭辉
廖伟霞
LI Haojun;WANG Xuhui;LIAO Weixia(Zhejiang University of Technology,Zhejiang Hangzhou 310023)
出处
《现代远距离教育》
2023年第1期19-31,共13页
Modern Distance Education
基金
2020年国家自然科学基金面上项目“基于大规模在线学习资源的个性化课程内容重构技术研究及应用”(编号:62077043)
2022年度浙江省哲学社会科学规划交叉学科重点支持课题“基于深度知识追踪的在线学习资源联动推荐服务研究”(编号:22JCXK05Z)。
关键词
弹幕
情感分析
深度学习
在线教育
情感词典
Barrage
Sentiment Analysis
Deep Learning
Online Education
Sentiment Dictionary
