Research On Students’ Classroom Attention Evaluation Based On EEG Signals

Education plays an important role in shaping future generations.The rapid development of artificial intelligence(AI)has led to considerable changes in the field of education.Advanced artificial intelligence technology may replace traditional classroom practice and pay more attention to students ’ preferences.Attention is closely related to students ’ learning efficiency.Evaluating students ’ attention is the cornerstone of improving teaching and learning.As an electrical signal that can reflect the individual ’s thinking state,EEG signals provide a new possibility for attention level recognition.Therefore,aiming at the related problems of students ’ classroom attention recognition and monitoring methods,system development and operation,this paper mainly completes the following work :Firstly,a simulation experiment of students ’ classroom behavior was designed and carried out.This paper takes the students of Hainan Normal University as the research object,and based on the ’ inverted U-curve ’ attention model theory,four attention-inducing experiments are designed to collect the individual EEG data of the subjects from high to low attention state respectively.Combined with the attention self-rating scale,the individual original EEG signal is screened and the data that meets the expected results of the experiment is retained.Finally,the individual EEG data of 32 subjects meet the expected results of the experiment and can be used for subsequent analysis and processing.Secondly,based on the theory of deep learning,a classification model of students ’classroom attention state is constructed.Firstly,FIR filter and ICA analysis are used to preprocess the original EEG data to remove noise interference and ocular artifacts.Then,the processed EEG data are downsampled and segmented.Then,the EEG feature extraction module is constructed with the CNN network as the main framework,which realizes the automatic extraction function of EEG features,and combines the Softmax layer and the fully connected layer to classify the EEG features.At the same time,in order to make full use of the temporal features in EEG signals,this paper also introduces the Transfomer model based on attention mechanism.The advantage of this network is that it completely abandons the traditional RNN network structure.The 20-fold cross-validation method was used to evaluate the performance of the classification model.The accuracy of the student classroom attention recognition and classification model based on the Transfomer attention mechanism was significantly higher than that of the CNN network classification model,reaching 87.9 %.Finally,based on Python language and Py Qt toolkit,a student classroom attention recognition and monitoring system based on EEG data is developed.The system realizes the real-time reading of EEG data and the effective identification and monitoring of students ’classroom status by means of data playback,and feeds back the test results to students,teachers and parents in time.