语言智能场景下在线课程学习行为情感语义分析与效果评价

Emotional Semantic Analysis and Effect Evaluation of Online Course Learning Behaviors in Language Intelligence Scenarios

利用语言智能技术解决在线课程学生学习行为情感语义分析存在的数据量大、分析耗费时间多、处理繁杂等问题,有助于实时掌握学习效果、改善教学方式。基于此,文章提出一种基于语言智能场景的学习行为情感语义分析ESAM-LI方法,该方法在梳理学习行为情感语义类型的基础上,首先获取在线课程学习行为文本信息并以基于Albert的微调模型方式进行模型训练;然后结合文本语言处理的BiLSTM模型和条件随机场CRF模型构建学习行为情感语义分析专业领域模型,同时构建情感信息标注标准并识别课程及教学知识点相关实体,获取学习行为情感语义极性类型;最后结合课程实体和教学知识点实体情感语义类型建立教学效果满意度评价模型并开展实证研究。实证效果表明,在课程效果评价和教学知识点评价上,ESAM-LI方法与传统教学效果评价方法两者高度一致,但ESAM-LI具有传统方法难以企及的处理速度,且能够克服人工主观误差,可为改善在线课程教学效果提供重要的借鉴。

Using language intelligence technology to address the issues of large data volume,time-consuming analysis,and complex processing in emotional semantic analysis of students’learning behaviors in online courses is of great significance for real-time monitoring of learning effects and improving teaching methods.In this regard,this paper proposed an emotional semantic analysis method based on language intelligence scenarios for learning.Based on sorting out the emotional semantic types of learning behaviors,this method firstly obtained the text information of learning behavior of online courses and carried out the model training by Albert-fine-tune.Subsequently,a specialized model for emotional semantic analysis of learning behaviors was cibstructed by combining the bi-directional long-short memory(BiLSTM)model of text language processing with the conditional random field(CRF)model.At the sam time,emotional information annotation standards were established and entities related to courses and teaching knowledge points were identified to obtain polar type of emotional semantic of learning behaviors.Finally,the evaluation model of teaching effect satisfaction was established by combining the emotional semantic types of the course entity and the teaching knowledge point entity,and further the empirical research was carried out.The empirical results showed that the ESAM-LI method was highly consistent with traditional teaching effectiveness in terms of the evaluations of course effectiveness and teaching knowldege points,but it outperformed traditional methods in terms of processing speed,and it can overcome the subjective errors in manual evaluation,which had important reference significance for improving the teaching effect of online courses.