多模态协作学习分析理论模型、实践逻辑和教育价值

The Theoretical Model,Practice Logic and Educational Value of Multimodal Collaborative Learning Analysis

涉及多主体、多因素、多场景的大规模多群组协作学习,正成为混合式学习的主流方式。深入挖掘协作学习中个体与群体的行为互动等外显性特征,与认知情感等内隐性指标的作用关系和演化机理,是实现“规模教育中的个性化培养”的关键。为此,遵循“理论、数据、分析、干预”的学习分析逻辑,构建了涵盖“行为-交互-认知-情感”的个体-群体双螺旋群体协作学习分析理论模型,随后,提出多模态数据采集、表征和跨域融合方法,并以分类和聚类为例展示了多模态关联分析方法和流程,介绍了分析结果可视化方法和画像技术,提出数据智慧支持的教育规律挖掘和干预机制,以期为智能教育服务和精准学习干预提供科学依据。

Large scale multi group collaborative learning is becoming the mainstream of blended learning,which involves multiple agents,multiple factors and multiple scenarios.Exploring the relationship between explicit characteristics such as individual and group behavior interaction and implicit indicators such as cognitive and emotion in collaborative learning is fundamental to realize“individualized cultivation in scale education”.Based on“theory,data,analysis and intervention”learning analytic logic,we constructed an individual-group double helix group collaborative learning analysis theoretical model,covering the four aspects of“behavior,interaction,cognition and emotion”.After that,we introduced multimodal data collection,representation and cross domain fusion methods,and took classification and clustering as examples to demonstrate multimodal association analysis methods and procedures,and introduced the visualization methods and portrait technology.Finally,we put forward an educational law mining and intervention mechanism supported by data intelligence to provide scientific intelligent education services and precision learning intervention.