如何帮助学习者走出学习资源迷航——基于学习者画像的个性化学习资源推荐  

How to Help Learners Get Out of Learning Resource Maze--Personalized Learning Resource Recommendations Based on Learner Persona

作　　者：成亚玲[1] 谭爱平[1] CHENG Yaling;TAN Aiping(Hunan Industry Polytechnic,Changsha Hunan 410208)

机构地区：[1]湖南工业职业技术学院,湖南长沙410208

出　　处：《当代职业教育》2023年第2期103-112,共10页Contemporary Vocational Education

基　　金：2019年湖南省“十三五”教育科学规划课题“人工智能支持下的个性化学习资源推荐与学习路径规划研究”(编号:XJK19BXX007)。

摘　　要：以MOOCs为代表的在线课程为学习者提供了多模态海量的学习资源,但学习者在选择资源时也经常面临信息过载、资源内容与学习需求不匹配等问题,造成学习资源迷航的困境。如何高效精准地向学习者推荐合适的学习资源与自适应的学习路径,已成为当前教育信息化亟待解决的问题。基于学习者画像的个性化学习资源推荐模型,从个性化资源推荐的现实需求出发,以现代教育理论和数字技术为支撑,探究个性化学习资源推荐过程中,学习者的状态与学习需求等个性化特征诊断、学习资源自身属性特征参数挖掘与表征、学习资源与学习需求精准匹配等问题,旨在向学习个体和学习群体推荐个性化学习资源。将基于学习者画像的个性化学习资源推荐模型应用于教学实践,结果表明,与基于协同过滤推荐、DINA认知诊断推荐相比,该学习资源推荐模型具有更高的精准度和更优的可解释性,能够为不同学习需求的学习者推荐适切的学习资源;同时该模型也在一定程度上缓解了冷启动与数据稀疏性等问题。Online courses represented by MOOCs provide learners with multimodal and massive learning resources.However,learners are often faced with information overload,and mismatch between resource content and learning needs when selecting resources,creating a learning resource maze.How to efficiently and accurately recommend appropriate learning resources and adaptive learning paths to learners has become an urgent problem in current education informatization.The personalized learning resource recommendation model based on learners'persona,starting from the practical needs of personalized learning resource recommendation,and supported by modern education theory and digital technology,explores the personalized feature diagnosis of learners'status and learning needs,the mining and representation of learning resources'own attribute feature parameters,and the precise matching of learning resources and learning needs in the process of personalized learning resource recommendation.It aims to recommend personalized learning resources to learning individuals and groups.The personalized learning resource recommendation model based on learner persona is applied to teaching practice.The results show that compared with collaborative filtering recommendation and DINA cognitive diagnosis recommendation,the learning resource recommendation model has higher accuracy and better interpretability,and can recommend appropriate learning resources for learners with different learning needs.At the same time,the model also alleviates the problems of cold start and data sparsity to some extent.

关 键 词：在线开放课程 学习者画像 学习资源 深度学习 推荐模型 数字化 

分 类 号：G720[文化科学—成人教育学]