Research On Online Learning Behavior Feature Mining And Academic Early Warning

The rapid development of Internet technology promotes the wide application of information technology in education and teaching.With the wide use of online learning system,people gradually accept and adapt to the way of online learning.Learners discuss and cooperate on various online platforms,learn courseware or take tests,resulting in a large amount of learning activities and behavior data.Because the data comes from different online learning platforms,there are some problems,such as scattered structure,inconsistent format and so on.In order to find valuable information from the original data and apply it,data mining related methods are needed to process and analyze the data.This thesis introduces the current situation of the application of data mining in the field of education,as well as the related research on the analysis of online learning behavior,and then expounds the clustering method,prediction method and academic early warning mechanism.This thesis uses data mining related methods to mine and analyze learners’ online learning behavior.By mining meaningful behavior characteristics for cluster analysis,this thesis explores the learning characteristics of learners,and establishes a learning early warning model to predict the risk of learners’ learning.The research results can provide high-quality decision-making basis for online learners,teaching workers and online teaching platforms to optimize teaching.Firstly,this thesis constructs an online learning behavior model,explores the correlation between online learning behavior characteristics and learning effect,and excavates meaningful behavior characteristics.Secondly,we make cluster analysis on learners,visualizes the clustering results,discusses the learning characteristics of different types of learners,and gives learners personalized learning suggestions.Finally,we construct the academic early warning model of online learning,determines the classification of academic early warning level,and provides suggestions for learners with different early warning levels.For the combination of multi-dimensional behavior characteristics,this thesis uses common classification algorithms to predict academic risk.Through the evaluation and analysis of the prediction results,we get the combination of behavior characteristics with better risk prediction effect and the classification algorithm suitable for early warning model.In general,this thesis analyzes the learning characteristics of different types of learners by mining the behavior characteristics of online learning.And we build an online learning academic early warning model to evaluate the failure prediction results of different combinations of behavioral characteristics.It is hoped that the research conclusion of this thesis can provide applicable suggestions for the improvement of learners’ learning effect,the formulation of teachers’ teaching plan,the optimization of online learning platform and so on.