基于教师近场语音的课堂教学过程自动分析

Automatic Analysis of Classroom Teaching Process based on Teachers’Near-field Speech

课堂教学过程分析自动化是改进日常教学的关键。随着AI技术的快速发展,使用商用软件对教师近场语音进行自动语音识别,能够获得精确的教师说话时段,得以开展如下的课堂教学过程自动编码与分析研究:(1)使用S-T(学生-教师)行为编码、构建S-T互动编码,实现了基于教师说话时段的全自动变时长S-T双编码;(2)提出了基于时长的S-T分析方法,生成教师主导行为占比Rt和师生互动率Ch的时间变化曲线;(3)通过教学活动建模,将连续的课堂教学过程离散为讲授、自主学习、问答-短互动、讨论-长互动等四类教学活动的序列;(4)开发了一套全自动可视化的课堂教学过程分析工具软件。

The automation of classroom teaching process analysis is the key to improve daily teaching.With the rapid development of artificial intelligence(AI)technology,using the commercial software for automatic speech recognition of teachers’near-field speech can obtain accurate teachers’speaking periods,and the following automatic coding and analysis research of classroom teaching process can be carried out:(1)using S-T(student-teacher)behavior coding and constructing S-T interactive coding can realize fully automatic variable length S-T double coding based on the teachers’speaking periods;(2)a duration-based S-T analysis method was proposed to generate the time variation curves of the proportion of teachers’dominant behavior Rt and the teacher-student interaction rate Ch;(3)the continuous classroom teaching process was divided into the sequences of four kinds of activities including lecture,autonomous learning,question answering and short interaction,discussion and long interaction through the modeling of teaching activities;(4)an automatic and visual tool software was developed for classroom teaching process analysis.