Exploring Online Learner Behavior Based On Educational Data Mining And Machine Learning

The online learning mode is very different from the traditional offline teaching method.The classic class-teaching mode has a relatively mature evaluation and early warning mechanism.Educators can gain insight into students’ learning status at any time and adjust teaching methods in a timely manner.For possible abnormal learning behaviors,they are also fully equipped to intervene and guide in advance.The online learning model is at a natural disadvantage in this regard,and online educators often lag behind in the grasp of students’ learning effects.With the extensive application of data mining technology and machine learning algorithms in the field of education,various studies based on online learning data have received more and more attention.This paper focuses on the learning behavior of learners in online learning platforms and applies data mining techniques and machine learning algorithms to the online education industry in an attempt to discover potential connections between different courses and universal learning patterns of learners,and to provide theoretical support and decision-making references for different participants in education and teaching activities.Through the comprehensive evaluation of the students’ course selection and learning enthusiasm,it is possible to understand the course tendencies and subject popularity of online learners.Using the association rule model based on the Apriori algorithm to build a course recommendation system for the online learning platform,we found that computer science occupies the absolute hotness among online courses and the extremely high lift proves that the system has a significant effect on the increase of users’ course registration.Using the time data of users watching course videos,we analyzed the learning activity and learning pattern of online learners and found that there is an obvious periodicity in their online learning time.In order to understand the long-term learning trend of online learners,the continuity of learners is reflected by the withdrawal behavior.With the help of single machine learning algorithms such as K-nearest neighbors,decision trees,logistic regression,and ensemble learning methods such as Bagging,random forest,XGBOOST,LightGBM,an online learner dropout prediction model is constructed based on the perspective of video playback.The results of the study demonstrate the feasibility of using video playback data to warn about class withdrawal behavior and the superiority of ensemble learning to fit this data.