【作 者】：
吴 娟， 陈 蕊， 吴 磊， 陈鹏鹤
【关 键 词 】：
阅读投入； 多模态数据； 混合场景； 阅读投入预警
【栏 目】：
学习环境与资源
【中文摘要】：
阅读投入是衡量中小学生在多种阅读情境中的参与水平与努力程度的有效指标，基于多模态数据对其进行精准表征与适时预警是促进良好阅读习惯养成、提升阅读素养的重要途径。鉴于此，研究明确了阅读投入的内涵与情感、认知、行为、社交四个构成维度，构建出混合场景下的阅读投入模型，并基于阅读过程中生成的文本、日志、图像、语音、生理、心理六种模态数据，形成了多模态数据表征阅读投入的框架与指标。以此为基础，阐述了不良阅读投入状态的预警路径，包括阅读投入的指标融合计算、状态识别分析、预警策略匹配和预警信息反馈。由此，可对中小学生阅读投入情况加以检视，及时发现学生的不良阅读状态并提供预警反馈，未来可通过阅读投入进一步揭示阅读理解能力的发展机制，并提供个性化干预策略。
【英文摘要】：
Reading engagement is an effective indicator of primary and secondary school students' participation and exertion in various reading contexts. Accurate characterization and timely warning of reading engagement based on multimodal data are important ways to foster good reading habits and enhance reading literacy. In view of this, this study explicates the connotation of reading engagement and four components of emotional, cognitive, behavioral and social dimensions, and constructs a model of reading engagement in blended scenario. Then, a multimodal characterization framework and indexes are established based on the data generated in the reading process: the text, log, image, phonological, physiological and psychological data. Based on the multimodal characterization framework, an early warning path of low reading engagement status is elaborated, including the calculation of index fusion, status identification and analysis, early warning strategy matching, and early warning information feedback. As a result, the reading engagement of primary and secondary school students can be reviewed, and the poor reading status of students can be detected and early warning feedback can be provided in a timely manner. In the future, the development mechanism of reading comprehension can be further revealed through reading engagement, and personalized intervention strategies can be provided.