Behavior Pattern based Performance Evaluation in MOOCs

With the development of MOOCs (Massive Open Online Courses), more and more courses could be studied online. Now researchers are showing an increasing interest in the field of MOOCs including dropout prediction, cheating detection and achievement prediction. Previous studies on achievement prediction mainly focus on students’ video behaviors and forum behaviors, and few researchers have paid attention to how well they do their assignments. In this paper, we choose a C Programming course as an experimental subject, which involves 1528 students. This paper mainly focuses on the students’ programming assignment accomplish behaviors and compilation information of programming assignments. In this paper, feature sequences were extracted from the logs according to submission times, submission order and plagiarism. The experimental results showed that the students who did not pass the exam had obvious sequence pattern, but the students who passed the test did not have obvious sequence pattern. Then we extracted 23 features from the compile information of students’ programming assignments, and selected the most distinguishing features to predict the students’ performance. The experimental results shows that we could get the accuracy rate of 70.49% for predicting students’ performance.