Research On Modeling And Prediction Of Learner Learning Behavior Based On Online Learning Community

With the development of online course programs for higher education institutions,more and more non-traditional learners have better learning opportunities.In today's fast-paced social atmosphere,through the online learning community,learners can conduct learning activities anytime,anywhere,which allows them to make the most of their time.However,because it is difficult for learners to find their own shortcomings,it is difficult for teachers to understand the learners' learning be haviors in depth,and their learning behaviors and learning states cannot be predic ted in time.Therefore,the learning effect of learners is likely to be more effective.It can be seen that it is crucial to model and predict the learning behavior ana lysis of learners in online learning.Learners generate a large amount of complex behavioral data in the learning process of the online learning community,which can reflect the learner's learning from the side.Constructing a learning behavior model is a common method to analyze learners' learning behavior,which can reflect the learner's academic situation more deeply and discover the hidden problems in their learning.For learning behaviors that have not yet occurred,we can predict the learning situation of the learners in advance by predicting their learning behav iors,and provide early warning ability for teachers or managers before the learners engage in certain bad behavior habits.Interventions.On the basis of summarizing the relevant research,in order to better model the learner's behavior,this paper investigates the click stream data of the learner's learning behavior in the online learning community,after fully considering the cha racteristics of the data:(1)From online learning The learning behaviors of learners in the community are classified,and the hidden Markov model is constructed based on the behaviors and learning motivations of different types of learners.Then,according to the behavioral state transition of the same class of learners,Const ruct their behavioral state transformation matrix and construct a Markov model of behavioral state;explore the relationship between learning behavior,learning state and learning effect through different granularity behavior models;(2)for better Mastering the learner's learning situation,able to find problematic learners in time,and achieving the effect of understanding the data from the learner's learning to predicting before learning.Based on the learning behavior model,this paper is based.on different learning strategies.Learners of the situation,learners' learning thro ugh Markov chains and superimposed Markovchains Behavior and learning status are predicted;(3)Finally,through empirical analysis,the online course data set of the course named Chinese Course in the data warehouse of pslcdatashop.org is used to experiment with experimental data to complete the modeling and prediction of learner behavior.The experimental results of learning behavior modeling and prediction were analyzed from different situations,and the relationship between learning behavior,state and effect was discussed,and the feasibility and effectiveness of the proposed scheme were proved.