数字化课堂作为数字教育的重要发展方向，是学生学习与发展的“催化器”。数字化课堂不仅可以提供交互式和个性化的学习体验，还可以通过数据分析与反馈服务，实现学生参与度的可计算性和可解释性，从而持续提升课堂教学质量和效果。本研究聚焦课堂数字化转型，通过融合情感计算和大数据分析等方法，构建数字化课堂环境下学习情感分析技术教育应用模型和学生参与度智能测评的分析框架，开展数字化课堂学生参与度测评的数据建模，研究数字化课堂环境下学生参与度的影响机制，从而提出数字化课堂环境下学习情感分析技术教育应用的治理建议，形成关于数字化课堂学生参与度的“理论研究-数据建模-实证研究-治理建议”全链条闭环式的研究体系。本研究共包含五个部分：

研究一和研究二均为理论研究，其中研究一通过研究数字化课堂中多模态学习情感分析的关键问题、逻辑理路与实施路线，构建多空间融合视域下多模态情感分析的理论框架，探索基于多模态数据融合的学生学习情感识别与智能评估，以揭示学生在数字化课堂视域下的学生情感发展机制。

研究二针对人工智能算法应用于学习行为和学生参与度的分析面临着可解释性不足，通用性欠缺、多模态融合弱和场景性欠缺等问题，基于四大基本设计理念，构建了数字化课堂学生参与度多模态自动化测评分析框架。

研究三是数据建模部分，基于数字化课堂中多模态学习情感分析技术应用模型和学生参与度多模态自动化测评分析框架，本研究建立了数字化课堂师生多模态数据采集解决方案，研发了在线众包标注平台CIT Label，建设了多模态课堂教学数据集，开发了学生参与度智能测评原型系统，实现了“师生基本动作-教学活动-群体学习场景-学生参与度”的智能识别与分析。

研究四是实证研究部分，本研究以968名中小学生为被试，采用问卷调查法对教师支持策略、情感氛围、技术接受度及学生参与度进行分析，探讨了教师支持策略与中小学生参与度的关系及其内在作用机制，重点考察了情感氛围在二者关系中的中介作用以及技术接受度的调节作用。研究结果显示：1）在控制班级排名、数字化课堂使用时长后，教师支持策略对学生参与度具有显著的正向预测作用；2）教师支持策略能够通过学生情感氛围的中介作用预测学生参与度；3）技术接受度不仅能够在教师支持策略与学生参与度之间的关系中起调节作用，而且能够对“教师支持策略-情感氛围-学生参与度”这一中介链条起调节作用。

研究五是治理建议部分，重点分析了教育领域中情感计算技术的价值和应用场景，探讨情感计算教育应用的困境与隐忧，并从内生性逻辑和关系性逻辑出发审视教育领域中情感计算应用隐忧的生成机制。基于此，提出四点实践规约：1）聚焦多学科交叉融合，攻克教育中情感计算的关键技术；2）制定情感计算技术审查标准与规范，实现跨人机的情感规则形塑；3）控制学生的隐私边界渗透，构建数字全景敞视下的多元协同共治；4）加强情感计算技术主体责任意识教育，提升情感计算教育应用生态效度。

总而言之，智能技术为学生参与度多模态自动化测评提供了一个有效且可靠的研究视角，也为改善数字化课堂教学质量带来了可能性。首先，本研究重点突出了全过程的多模态分析与场景自动融入，构建了数字化课堂学生参与度智能测评理论体系。其次，基于此，研发了面向数字化课堂的师生多模态数据采集解决方案、在线众包标注平台CIT Label、多模态课堂教学数据集、学生参与度智能测评原型系统等系列支撑工具，完成了数字化课堂学生参与度测评的数据建模和实证研究。最后，开展智能技术治理研究，提出了适用性强且指导性高的实践规约与研究建议，为我国情感计算技术规范而有序地融入教育生态提供了理论借鉴和行动指南。

As an important development direction of digital education, digital classroom is a "catalyst" for improving student engagement. Digital classroom can not only provide interactive and personalized learning experiences, but also achieve calculability and interpretability of student engagement through data analysis and feedback services, continuously improving the quality and effectiveness of classroom teaching. This research focuses on the digital transformation of the classroom. By integrating methods such as affective computing and big data analysis, we built the educational application model of learning affective analysis and the analysis framework for assessment of student engagement in a digital classroom. We carried out data construction for student engagement assessment in the digital classrooms. We surveyed the influencing mechanism of student engagement in the digital classroom, proposing governance suggestions for the educational application of learning effectiveness analysis technology. So we formed a "theoretical research-data modeling-empirical research-governance suggestions "full-chain closed-loop research system on student engagement in the digital classroom. This research contains five parts:

Study 1 and Study 2 were both theoretical research. Study 1 constructed a theoretical framework for multi-modal effectiveness analysis from a multi-space fusion perspective by studying the key issues, logic and implementation routes of multi-modal learning effectiveness analysis in the digital classroom. Explore student learning effectiveness recognition and intelligent assessment based on multi-modal data fusion to reveal students' emotional development mechanism in the digital classroom.

Study 2 aimed at the analysis of artificial intelligence algorithms applied to learning behavior and student engagement, which faced problems such as insufficient interpretability, lack of versatility, weak multi-modal integration, and lack of scenario. Based on four basic design concepts, it was constructed the analysis framework for multi-modal automated assessment of student engagement in the digital classroom.

Study 3 was the data modeling. Based on the application model of multi-modal learning effectiveness analysis and the multi-modal automated assessment framework of student engagement in the digital classroom, this study established a multi-modal data collection solution for teachers and students, and developed CIT Label, an online crowdsourcing annotation platform. We built a multi-modal classroom teaching data set, developed a prototype system for intelligent assessment of student engagement, and realized the recognition and analysis of "teachers and students' basic actions - teaching activities - learning scenarios - student engagement".

Study 4 was the empirical research. This study took 968 primary and secondary students as subjects, using a questionnaire survey to analyze teacher support strategies, emotional atmosphere, technology acceptance and student engagement, and explored the relationship between teacher support strategies and primary and secondary student engagement. We focused on the mediating role of emotional atmosphere in the relationship between the two and the moderating role of technology acceptance. The results of the study showed that: 1) After controlling class ranking and the duration of the digital classroom, teacher support strategies had a significant positive predictive effect on student engagement; 2) Teacher support strategies can predict student engagement through the mediating role of students' emotional atmosphere; 3) Technology acceptance can not only mediate the relationship between teacher support strategies and student engagement, but also mediate the intermediary chain of "teacher support strategies-emotional atmosphere-student engagement".

Study 5 was the governance suggestions, which focused on analyzing the value and application scenarios of affective computing technology in education, discussed the dilemmas of affective computing education applications, and examined the worries of affective computing applications from the perspective of endogenous logic and relational logic. generation mechanism. Based on this, four practical protocols were proposed: 1) Focus on cross-disciplinary integration and conquer the key technologies of affective computing in education; 2) Develop review standards and specifications for affective computing technology to achieve the shaping of emotional rules across humans and machines; 3) Control penetrating students' privacy boundaries and build diverse collaborative governance under the digital panorama; 4) Strengthen the education of subject responsibility awareness of affective computing technology and improve the ecological validity of affective computing education applications.

on the whole, intelligent technology provided an effective and reliable research perspective for multi-modal automated assessment of student engagement, and also brought the possibility of improving the teaching quality of digital classroom. First of all, this research built a theoretical system for the intelligent evaluation of student engagement in the digital classroom, focusing on the multi-modal analysis of the whole process and and the scene automatically blended into it. Secondly, based on this, a series of supporting tools such as a multi-modal data collection solution for teachers and students in the digital classroom, an online crowdsourcing annotation platform CIT Label, a multi-modal teaching data set, and a student engagement intelligent assessment prototype system were developed to realize the  data modeling and empirical research on assessment of student engagement. Finally, we carried out intelligent technology governance and put forward highly applicable and instructive suggestions, providing theoretical reference and action guidance for the standardized and orderly integration of affective computing technology into the educational ecosystem.