【作 者】：
田 阳， 陈 鹏， 黄荣怀， 曾海军
【关 键 词 】：
多模态； 混合学习； 多模态交互分析； 学习分析； 优化策略
【栏 目】：
课程与教学
【中文摘要】：
“互联网+教育”促进了混合学习的实践应用。混合学习融合了物理空间、社交空间、资源空间等，这使得学习交互的内在机理成为学术界重点关注的议题之一，研究团队尝试从混合学习的多模态分析视角揭示混合学习交互内在机理。通过对混合学习的多模态交互特征分析，归纳出其特征为跨空间交互、同步与异步交互并存、多情境转换、具身情感性、可表征等。结合混合学习运行规律，发现混合学习的多模态交互过程中可表征多模态和抽取多模态信息，并总结出多模态交互分析的主要流程为数据采集、加工与筛选、数据分析、数据表征可视化、反馈调节等。多模态交互信息经过可视化表征，将分析得到的结果经过仪表盘反馈给教师、学生，帮助教师和学生及时调整教与学的方法，并促进混合学习多模态交互策略的优化。
【英文摘要】：
"Internet + education" promotes the practical application of blended learning. Blended learning integrates physical space, social space, resource space, etc., which makes the internal mechanism of learning interaction become one of the main topics in academia. This research team tries to reveal the intrinsic mechanism of blended learning interaction from the perspective of multimodal analysis. Through the analysis of the multimodal interaction features of blended learning, the features are summarized as cross-space interaction, synchronous and asynchronous interaction, multi-situational transformation, physical emotion, characterization and so on. Combined with the rules of blended learning, it is found that in the multimodal interaction of blended learning, the multimodal information can be represented and extracted. Moreover, the main processes of multimodal interaction analysis are data acquisition, processing and screening, data analysis, data characterization visualization, feedback adjustment, etc. After the multimodal interaction information is visualized, the results are fed back to teachers and students through the dashboard to help them adjust teaching and learning methods in time and promote the optimization strategy of multimodal interaction in blended learning.