體驗區
|
| 篇名 |
基于人工智能引擎自动标注的课堂教学行为分析
|
| 並列篇名 |
Automated Annotation of Classroom Behaviours with Al Engine
|
| 作者 |
卢国庆(LU Guoqing)
、谢魁(XIE Kui)
、刘清堂(LIU Qingtang)
、张臣文(ZHANG Chenwen)
、于爽(YU Shuang)
|
| 中文摘要 |
课堂教学行为是影响课堂教学效果的重要因素，已有的课堂教学行为采集存在劳动密集、分类模糊和编码复杂等不足。人工智能技术为课堂教学行为大数据伴随式采集、自动化智能标注提供了新的契机。本研究以西北地区某市三所学校的1201个常规课堂教学视频为研究样本，利用人工智能引擎自动标注课堂教学行为，并采用相关性分析、主成分分析、非参数差异性检验等方法，对课堂教学行为类型、规律及差异性进行分析。研究发现：1)课堂教学各类行为的出现频率不等且相差较大，其中，读写、讲授、巡视行为占比较大，生生互动、师生互动占比较小；2)多数课堂教学行为之间具有关联性，其中，应答与生生互动之间的相关性最高，巡视、读写与其他行为存在负相关；3)教师行为和学生行为并非完全属于不同的成分；4)不同特征教师的课堂教学行为之间存在差异；5)不同类型课堂的教学行为之间存在差异。研究结论可为人工智能时代挖掘课堂教学行为规律、改进课堂教学及开展教研活动提供参考。
|
| 英文摘要 |
Classroom instructional behavior is an important factor that affects instruction and learning, while the existing instructional behavior labeling has some problems, such as labor-intensive, fuzzy classification, and complex coding. Artificial intelligence (AI) technology provides new opportunities for seamless and anytime data collection of instructional behaviors. Based on sorting out the evolution of instructional behavior labeling, this study focuses on exploring the characteristics of large-scale classroom instructional behavior and the difference of classroom instructional behavior in different teachers and instructional models. Research samples included 1201 classroom videos from three junior high schools in the northwestern region of China. An AI engine was used to automatically mark teaching and learning behaviors, and obtain a large sample of classroom behavior data. Then, correlation analysis, principal component analysis, and difference analysis methods were used to analyze the characteristics, type, and difference of classroom behaviors. The study found that: 1) The frequency of instructional behaviors varied widely. Among them, reading-writing, lecturing, and inspection behaviors accounted for a large proportion of all behaviors. On the other hand, there were relatively fewer student-student and teacher-student interactions; 2) Most behaviors related to each other. For example, the student response and student-student interaction had the highest correlation score. However, teacher inspection, reading-writing, and other behaviors were negatively correlated; 3) Two factors were extracted through principal component analysis, and teacher behaviors and student behaviors were not completely different components; 4) Teachers with different characteristics showed differences in their instructional behaviors; 5) There were differences between instructional behaviors in different instructional models. The research provides a method and practical reference for finding the characteristic of instructional behavior, improving classroom instruction, and carrying out teaching and research activities in the era of AI.
|
| 頁次 |
097-104
|
| 關鍵詞 |
人工智能
、课堂教学行为
、智能标注
、相关性分析
、差异性分析
、artificial intelligence
、instructional behavior
、smart annotation
、correlation analysis
、difference analysis
、
|
| 卷期 |
27:6
|
| 日期 |
202112
|
| 刊名 |
開放教育研究
|
| 出版單位 |
上海遠程教育集團、上海電視大學
|
| DOI |
10.13966/j.cnki.kfjyyj.2021.06.011
|
|
相關文獻 :
客服信箱