Online Student Learning Engagement Measurement Incorporating Multidimensional Data

Thanks to the rapid development of information technology such as artificial intelligence and multimedia,as well as the practical needs of modern education,online education has become a very important part of the current diversified education system.Although online education can maximize the use of resources and break through the limitations of teaching time and space,the inability to effectively supervise students makes students’ learning engagement lower and learning effectiveness poorer.Student learning engagement is considered a prerequisite for online learning and plays a key role in student course persistence and learning outcomes.Therefore,in online education,how to make good use of the teaching platform and keep track of students’ learning engagement is important to ensure teaching quality and improve students’ learning effectiveness.In this paper,we analyze the objective data generated in the process of online learning to measure students’ learning engagement from multiple dimensions,and the main research contents are as follows.(1)Construct a four-dimensional model of learning engagement.The four-dimensional model of learning input was constructed from four dimensions: behavioral input,social input,emotional input and cognitive input,and finally 14 evaluation indexes were selected to build an online learning input evaluation index system by combining the characteristics of students’ online learning activities.(2)Data collection and processing.In online learning,all learning activities of learners are carried out on the online learning platform and recorded in real time.Statistics and analysis of log data,interaction data,and test scores are conducted separately,and students’ behavioral input,social input,and cognitive input can be understood.Since emotional input belongs to the implicit state of learners,the log and interaction data have limited ability to characterize it.Therefore,in this paper,we analyze emotions by acquiring students’ classroom expression data with the camera of cell phones or computers.(3)Emotional engagement measurement of students based on expression recognition.LBP,HOG and the fused facial expression feature extraction method combined with SVM expression classification method were applied to the collected expression dataset respectively,and the results showed that the overall recognition rate of the extraction algorithm based on the fusion of LBP and HOG features reached the highest 92.3%,and the lowest overall recognition rate was 86.4% for HOG only.Therefore,this paper selects the extraction algorithm based on LBP and HOG feature fusion for expression recognition,and scores the expressions according to the different effects of classroom expressions on learning,so as to realize the quantitative representation of students’ classroom emotions.(4)Learning input analysis based on hierarchical analysis and BP neural network.A hierarchical analysis was used to determine the weights of each evaluation index,and the results of the subjective weighting were used to "guide" the training of the neural network,and a BP neural network model was constructed on this basis for evaluation.The results show that the method performs well on the test set data.