Research And Implementation Of Dynamic Monitoring And Visualization Of Online Learning Behavior

In recent years,with the rapid development of Internet technology and the continuous improvement of social demand,online education has become a more and more popular education mode.In this kind of education mode,teachers upload learning materials,students arrange learning time reasonably according to their own actual situation,and choose learning environment and learning content independently,which effectively solves the problem of time and space limitation in traditional classroom.But there are also some shortcomings,such as teachers can not observe students' learning state and evaluate their learning performance in time,students do not have the ability to objectively evaluate their learning state and lack of incentive mechanism for independent learning.Therefore,it is of great practical significance to dynamically monitor and analyze students' learning behaviors in the process of online learning and to reflect students' learning status in time and objectively.This paper analyzed the online learning behavior of the registered users,Based on the online learning platform of College of Network and Continuing Education of Xi Dian University.In this paper,takes database server and web log server of online learning platform as data source.Build a four-tier online learning behavior data model,conduct statistical analysis on the behavior characteristics data,and use spss20.0 to analyze the correlation between the behavior data and academic record.The K-Means algorithm is used to cluster the online learning behavior.The learning state is divided to realize the short-term monitoring and evaluation effect of online learning behavior.The academic record prediction model based on BP neural network is constructed to realize the long-term monitoring effect of online learning behavior.The analysis results are fed back by data visualization technology.In view of the application of K-Means algorithm in learning state division,this paper considers the factors,such as the large number of learning behavior features and the unclear classification of learning state division.Using elbow method to determine the number of the best clustering centers,and then determine the classification of learning state.At the same time,to solve the problem that k-means algorithm is sensitive to the initial clustering center,this paper uses the farthest strategy to select the clustering center.The experimental results show that the improved algorithm is more efficient.Based on the online learning behavior data after the above analysis and processing,the academic record model based on BP neural network is established.In order to improve the prediction accuracy and generalization ability of the academic record model,this paper introduces Ada Boost algorithm which is combined with BP neural network to optimize the performance prediction model.At the same time,a comparative experiment is established to compare the mean square error of the model before and after optimization and the state diagram of the prediction results.The results show that the output of the academic record prediction model based on Ada Boost-BP neural network is relatively closer to the expected value.Finally,based on the online learning behavior state division clustering model and academic record prediction model,combined with data visualization technology,the online learning behavior monitoring system is constructed,The online learning behavior monitoring results are visualized according to the learning behavior analysis module and performance prediction module.