The Application of CART Algorithm in Analyzing Relationship of MOOC Learning Behavior and Grades

Nowadays, with the coming of education informatization, MOOC is developing vigorously. Accumulating a large number of students' behavior data in the online teaching platforms has been widely concerned. In this paper, using the CART algorithm of decision tree, we analyze the data of MOOC about medical science in Stanford University. And then we explore the factors and importance influencing the test results. The study found that the total number of questions, the score of assignments, and the number of quizzes could be used as an important indicator of performance prediction. We also found that the decision tree model constructed by this course is as accurate as 90%, and the value of CART algorithm for online learning effect is tested. Finally, based on the related research of Stanford University, the paper puts forward some suggestions to improve the service mode of MOOC platforms.