Research On The Analysis Method Of Teacher’s Classroom Question Based On Deep Learning

Nowadays,Artificial intelligence technology is widely used in classroom teaching for teachers to provide accurate teaching reflection,greatly assist teachers effective classroom teaching activities,support teacher’s professional ability development,but also requires teachers to learners and teaching contents and teaching methods have a more clear understanding,to be able to clear the needs of learning and knowledge level,from the classroom efficiency questions,reasonable planning,reflection and adjust the learners’ learning process,so as to achieve the aim of classroom teaching.Classroom questioning is the hinge component of teachers’ classroom teaching behavior and the principle way for teachers and students to mutual interact in classroom.The analysis of teachers’ classroom questions is the critical to improve the standard of teachers’ classroom teaching.In order to facilitate the development of teachers’ professional ability perfect and enable them to effectively reflect on classroom questioning so as to indirectly improve the development of learners’ thinking ability,this research aims to automatically analyze classroom teachers’ questioning discourse,and adopts deep learning algorithm to automatically classify classroom questioning discourse in real classroom teachers’ teaching.Before all else,Based on the existing research on classroom questioning,the classification framework of teachers’ classroom questioning is designed with the learner as the center,and the classification ultimately points to the improvement of learners’ advanced thinking ability.Therefore,combining with the existing classification of classroom questioning and the characteristics of real classroom questioning corpus,this paper gives an analytical framework from two dimensions of questioning content and type.Secondary,Based on the existing research and the designed classification framework of classroom questions,a class question classification model based on deep learning is constructed.In this study,CNN model and LSTM model were selected as the classifier model of classroom question text,and a series of experimental parameters were set.By constructing the classification model of classroom questioning,the classification results of questioning are fed back to the teachers,in order to provide the classroom questioning discourse reference for the teachers and direct the teachers to change the teaching questioning strategies.The model of teacher questioning level is conductive to help novice teachers grow into expert teachers,so as to indirectly improve the development of learners’ thinking ability.At length,in this study,convolutional neural network(CNN)and short and long time memory network(LSTM)were used to classify 9090 texts of teachers’ questions from 80 classes.By distinguishing the data sets,the results show that the CNN model has the best classification effect,and its overall accuracy in the two dimensions of question content and question type is 85.17% and 87.84%,respectively.At the same time,from the change curve of accuracy and Loss value of CNN model in training set and verification set,it can be seen that in the process of training data set of CNN model.The model is stable after the accuracy rate increases sharply,and the Loss value is stable after the Loss value shrinks sharply.The training process of the model performs well.