A Hybrid Approach for Dropout Prediction of MOOC Students using Machine Learning

Massive Open Online Courses (MOOC) is an extensive way of providing online education to the students all over the world. Based on the statistics, this education system have millions of students attending hundreds of courses in different offered programs. Since, MOOC started, it has been facing a challenging concerns, which is also a major difference between the traditional teaching and MOOC, known as "student dropout ratio". With this fact, the overall performance of MOOC is negatively impacted the real purpose of distance learning. Whereas the difference between course registration and course completion ratio in MOOC is quite large. On the better side, the emerging technologies has created several opportunities for the students to get education online, but due to multiple factors the dropout ratio of online students is high as compare to traditional school learning process. This research is focusing on the issue to understand and predict the MOOC dropout ratio. The multiple models and evaluation metrics generating variety of results as extracted from literature review. To tackle this problem, the experiment conducted in this study using KDD MOOC dataset by implementing hybrid approach of machine learning algorithms. The results suggested the appropriate improvements in the dropout accuracy ratio. Based on the fmal results, the maximum accuracy recorded as 90% that measured through random forest model. Finally, the model can help and assist the online education system to understand the early dropout prediction and to do necessary arrangements.