Learning Analytics in a Blended Computer Education Course

Abstract

Massive open online courses (MOOCs) have been widely used and many institutions have invested considerable effort in developing, promoting and delivering such courses. Online educational pattern, both inside and outside of the campus community, have been become more popular. In recent years, a new research agenda has emerged, focusing on predicting and explaining dropouts and MOOCs with low completion rates. However, due to different problem specifications and evaluation metrics, performing learning analytics (LA) of online education is a challenging task. The online learner has a variety of purposes, such as complete learning courses, some knowledge points in work-based learning courses, selective learning for reviewing exams, certificate-based learning, and watching videos for interest. According to users'online behavior data, it is difficult to study why so many people drop out of MOOCs. Because users have many reasons, such as not understanding, bad network connection, other things to do, and so on. This paper study was performed to analyze data of students' online activity in a blended computer education course in Northwestern Polytechnical University (NPU) to identify quantitative markers that correlate with students' performance and might be used as early warning signs for possible data-driven measures. We applied the research methodology to three online courses offered generously by the Chinese University MOOC (CUMOOC) to evaluate the effectiveness.