Research And System Implementation Of Learning Attention Recognition Based On PPG In MOOC

MOOC has broken the traditional offline teaching requirements for schools and teachers,so that more learners can obtain rich and high-quality online learning resources,which has become one of the important learning methods at present.However,unlike the traditional classroom teaching,teachers can understand the student's learning attention face to face.The asynchronous nature of MOOC teaching leads to feedback delays and lack of personalized guidance.How to effectively recognize the learning attention in the MOOC learning process in order to adjust the teaching content in time and provide adaptive feedback,so as to improve the learning effect of students has become an important factor affecting the participation of the MOOC.With the maturity of sensing technology,the collection of attention is also gradually breaking through the limitations of technology.The physiological data obtained based on wearable devices can quantify and understand the learning status of learners,providing an effective means for attention recognition in the MOOC environment.In this thesis,on the basis of fully investigating the research on learning attention recognition at home and abroad,I conducted an in-depth study on the non-invasive recognition and visual analysis of attention in the MOOC environment,researched and constructed a PPG-based learning attention recognition model and method,design and development The identification system was tested and verified.The main research results are as follows:First,this thesis proposes a learning attention recognition method based on PPG.By building a non-intrusive perception environment,setting reasonable learning tasks,and selecting wearable sensor devices,the photoelectric volume pulse wave data and learning attention information of the subjects when studying in the Mu class environment were effectively obtained.The method of sliding window and overlapping window is used to preprocess the data,extract the time domain,frequency domain and nonlinear features,construct the feature matrix,use the random forest algorithm to construct the recognition model,and use grid search to determine the hyperparameters of the classifier.The empirical test verifies the model's ability to recognize learning attention,and realizes the effective recognition of learning attention in the MOOC environment in a portable and non-invasive way.Secondly,this thesis builds the system architecture of attention recognition in the MOOC environment.In the form of a questionnaire survey,investigate the actual needs,opinions,and suggestions of MOOC's actual use groups for the introduction of learning attention indicators into MOOC.Combining the results of the questionnaire survey with the current status quo,the functional requirements and non-functional requirements of the system are analyzed from the perspective of the teacher and learner.Based on the attention level display method that conforms to the individual's cognitive law,taking into account data privacy and applicability,an effective presentation form of the attention state under the MOOC environment and a feedback form that effectively integrates and analyzes the learning attention indicator with the MOOC course itself are designed.Finally,integrated learning attention recognition model and system design,completed the development and functional test of the MOOC system including learning attention function adapted to multiple terminals,and conducted user experience,functional design,privacy protection,system cost,system technology,comparative evaluation of six dimensions of interface design.