Research On Online Learning Behavior Analysis And Learning Performance Prediction In Big Data Environment

The rapid development of Internet information technology and education informatization has changed traditional learning methods,making online learning methods rapidly emerging and widely used in human resource development and education.However,online learning methods also face some challenges.The separation of teaching and learning between teachers and students prevents teachers from observing the learning status and learning behavior of learners,unable to give learners personalized learning guidance,and unable to intervene learners in time Poor learning behavior.As a result,the learning effect of learners is poor,and the phenomenon of user loss in the platform occurs frequently.A large number of learners’ learning behavior data are recorded in the online learning platform system in the big data environment.By analyzing the learning behavior data,the learning behavior model of online learners is optimized,and the online learning platform is promoted for better development.On the basis of existing research,under the guidance of related theories and models,this thesis constructs an online learning behavior indicator system and analysis framework based on the relevant classification model of online learning behavior and the index characteristics of the open data set of the learning platform.On this basis,obtain online learners’ learning behavior data,and conduct research from three dimensions.First,understand the basic situation of learners and analyze whether there are differences between the learning behaviors of learners with different characteristics.Second,analyze the correlation between learning behavior and academic performance.On the basis of significant correlation,the learning performance prediction model is determined through multiple linear regression analysis.In order to verify the accuracy of the model and determine whether the learner’s academic performance has early warning risks.Use logistic regression and neural network models to classify and predict learning performance.Compare and analyze the advantages and disadvantages of the two models and the accuracy of prediction.Summarize the analysis results of online learners’ learning behavior,optimize online learner’s personalized learning model and online learner’s learning behavior intervention model.Through research,it is found that there are significant differences in the learning behavior of learners with different characteristics,and the results of the analysis of differences are different.There is a significant positive correlation between learning behavior and academic performance in all dimensions.Operational behavior and problem-solving behavior are highly correlated with academic performance.Logistic regression verifies the accuracy of the learning performance prediction model.Among them,"learning event span","module browsing ratio","posting number" and "module completion ratio" are the main learning behavior factors that affect learning performance.The online learner’s personalized learning model provides learners with personalized learning planning and guidance,and encourages learners to learn independently.The online learner’s learning behavior intervention model intervenes bad learning behaviors from the three perspectives of platform managers,teachers and learners to improve learners’ academic performance.