Research On Classroom Teaching Behavior And Teaching Content Summary Based On Teaching Video

The analysis of classroom teaching behavior and teaching content is the key link and main means to improve the quality of classroom teaching.One of the ways.In recent years,the rapid development of science and technology such as artificial intelligence and the advancement of informatization have also promoted the pace of educational technology reform,providing technical support for the use of more efficient and accurate methods to carry out research on classroom teaching behavior and teaching content.The main content of this research is the automatic recognition of classroom teaching behavior based on teaching video and the generation of knowledge map of teaching content.Combined with natural language processing,deep learning and other methods at this stage,a classroom teaching behavior framework suitable for automatic recognition is constructed based on teaching video.The classroom teaching behavior framework divides teaching behavior into two parts: classroom environment and classroom structure,and refines 22 behavioral indicators such as questioning,teaching,observation,and clear learning goals.Part of the behavioral indicators involving teachers’ discourse are identified by reference to the metadiscourse analysis method,and the teacher’s metadiscourse analysis method in the teaching scenario is proposed,the identification rules are formulated,and the teacher’s metadiscourse indicators of 18 classroom teaching videos in many primary and secondary schools in Beijing are completed.Automatic identification.The time distribution or location distribution characteristics of the identification results of each teaching behavior index in the 18 classrooms were analyzed one by one,and the accuracy of the identification results was tested in combination with the actual classroom situation in the teaching video.high accuracy.Integrate the classroom teaching behavior data of each class,use K-Means clustering to divide the 18 classes into three categories,and analyze the characteristics of various teaching behaviors in the clustering results.Interactive,teacher-led teaching.Based on the teacher’s speech in the classroom,the Text Rank keyword extraction algorithm is used to extract keywords with a rank greater than 0.1 as classroom teaching knowledge points,and the Own Think knowledge base is used to query the relationship between knowledge points.-Knowledge point" triple information is stored in the Neo4 j graph database and visualized to generate a knowledge graph of teaching content.By analyzing the relationship between the generated knowledge map and the actual teaching content,it is proposed to form a knowledge network from knowledge points,the relationship between knowledge points,and knowledge points and relationships.Whether the information in these three levels is relatively accurate,complete,and without redundancy The idea of evaluating the validity of the knowledge graph of the teaching content is based on the remaining information,and the validity of the knowledge graph is verified.The scoring results show that most of the teaching content knowledge graph generated based on the teacher’s words in the teaching video has a high score.In the two chemistry classes,there are many transliteration errors of professional terms in the audio transcription process,resulting in poor knowledge graph effect.Integrate the results of automatic recognition of classroom teaching behavior based on teaching videos and the knowledge map of classroom teaching content to form a classroom briefing report,which concisely summarizes and presents the information of teacher behavior and teaching content in the classroom.