Analysis and Forecasting of the Outflow of Students of Massive Open Online Courses

Massive open online courses (MOOCs) first appeared on the Coursera and Edx platforms in 2008 and became popular in 2012. In the eighth year of its existence, MOOCs brought together 110 million students from around the world and covered more than 900 universities. Despite the growing popularity of online courses, an urgent problem in the world of online education is that only a small proportion of students reach the end of the course. According to statistics from online learning platforms, approximately 90% -95% of students do not complete the online course. This work is devoted to the course completion share improvement. Using the Python system, the data of the online course “Introduction to Data Analysis in R” on the Russian educational platform Stepic were analyzed and predictive models were constructed using machine learning algorithms. Also given recommendations to the organizers of the courses to increase the share of students who completed the course. Keywords Massive open online courses, MOOCs, outflow rate, personal goals, MOOC efficiency, machine learning, predictive model, outflow prediction.