Research On Learning Behavior Modeling And Learning Effect Prediction Of MOOC Based On Social Cognition Theory

With the popularization of life-long education ideas and the deepening of the concept of learning-oriented society,more and more online learning platforms such as MOOCs are springing up in China.Although these online learning platforms are well-organized,resource-rich,and technically sophisticated,they still cannot avoid the common disadvantages of online learning—teachers and students.Because it is impossible to intuitively observe a series of learning behaviors of learners,benign interventions and reasonable guidances cannot play a role,and many learners end up with halfway or failure.The big data feature of education provides us with a solution to the problem.Learners generate a large amount of data each day on various learning platforms.By mining online learners' behavioral data,understanding leaner's behavior characteristics and learning patterns,and analyzing learning problems,predicting learning results is an effective way to shorten improving cycle of learning and enhance online learning effect in the era of big data.A large number of studies have shown that the online learning behavior of MOOC learners is an important variable for predicting learners' learning effects.With the deepening of research,people gradually realize that online learning behavior is affected by many factors,especially learners' learning psychology.Therefore,if we can not fully analyze the influencing factors of online learning behavior,we can not really explain the learning behavior of online learners,nor can we provide a basis for evaluation and prediction of online learning effects.Ithough the granularity of research in this area is continuously being refined,the main concern of all educators is still how to select factors that affect learning behavior,how to divide online learning behavior,how to effectively model learner learning behavior,and how Effectively apply data mining technology and learning analysis technology to the prediction of online learning effects.The social cognitive theory emphasizes that learning behavior is not only influenced by the external environment,but also by the individual's psychological characteristics.Based on this,this study divides the factors that affect learning behavior into learning environment factors and learning psychological factors.They are collectively referred to as learning situations,and discuss the relationship between learning situations,learning behaviors and learning effects.On the basis of summarizing related practice and theoretical research,this study first classifies the MOOC learning situation from two aspects of online learning environment and online learning psychology,and classifies the MOOC learning behavior from the perspective of cognitive participation in learning behavior.This establishes a MOOC learning behavior model based on social cognitive theory;then,through the analysis of real MOOC curriculum data,it constructs specific learning situation indicators and learning behavior indicators,and explores the complex relationship between learning situations,learning behaviors and learning effects.Amendments to the MOOC Learning Behavior Model based on this analysis.In addition,this study also verified the effectiveness of the model through logistic regression analysis and learning effect prediction research.The results show that the proposed MOOC learning behavior model based on social cognitive theory is more accurate in the analysis of the relationship between learning context,learning behavior,and learning environment,in addition,the deep neural factor decomposition model can handle complex feature interactions and improve the accuracy of learning predictions.