摘要
多模态学习分析研究为全面、精准洞悉学生学习过程,促进有效学习,实现精准干预提供了可能。通过对2017年至2021年学习分析与知识国际会议(LAK)中的多模态学习分析主题文章进行分析发现,当前研究者已借助行为、生理、心理、基本信息等不同模态数据围绕学习过程理解、学习评价、学习支持与反馈展开了探索,同时也就多模态数据的采集与处理等基本问题展开了探讨,但是当前研究在样本数量、研究持续开展时间以及不同模态数据的融合分析等方面还存在着不足。为促进多模态学习分析研究的开展,未来可基于学习终端及智慧学习环境实现多模态数据的采集,基于智能技术实现不同模态数据特征的提取与整合分析,加强跨学科交叉合作,同时也应关注多模态数据采集与应用的伦理与隐私保护问题。
Multimodal learning analytics(MMLA) provides an opportunity for understanding students’ learning process comprehensively and providing intervention accurately. After analyzing researches of multimodal learning analytics from the recent five International Conferences on Learning Analytics and Knowledge(LAK 2017-2021), it is found that previous studies have used different modal data to understand learners’ learning process, evaluate learners performance and provide supports and feedback, as well as discuss some issues about the multimodal data collection and processing. Nevertheless, there are also some deficiencies in the sample size, research duration, and fusion analysis of different modal data. In the future, it is suggested to collect multimodal data based on learning terminals and in smart learning environments, to extract and integrate features of different modal data based on intelligent technology, and to strengthen cross-disciplinary cooperation. Moreover, ethics and privacy in multimodal data collection and application is worth paying attention to.
作者
高明
王小霞
GAO Ming;WANG Xiaoxia(Faculty of Education,Beijing Normal University,Beijing 100875,China;Education Informatization Research Center,Shangrao Normal University,Shangrao 334001,China)
出处
《开放学习研究》
2022年第1期45-54,共10页
Journal of Open Learning
基金
2020年上饶师范学院校级自选课题“基于眼动追踪技术的在线测试中学习者眼动行为模型研究”(课题编号:202022)的研究成果。
