Massive Open Online Courses (MOOCs) are exclusively designed for virtual teaching resources of high quality that benefit large audiences. This paper explores the current status of meta-analysis on MOOCs. The key issues identified during the study in predicting MOOCs include characterizing the prediction learning outcomes, identifying the prediction features, and determining the methodologies utilized to forecast the variables. The assessment is one of the research efforts to collate the leading technologies and concepts used in learning to forecast the achievement of student learning outcomes as well as the leading characteristics utilized, which are a factor in educational outcomes. A wide range of prediction features are available, but video data and behavioral analysis of the platform are the most prominent ones. Our findings suggest a strong desire to predict MOOC dropouts in various learning approaches and two of the most commonly used forecasting analytics are Logistic Regression and Support Vector Machines.
Keywords: MOOC prediction, Learning Analytics, Performance Prediction, Learning Outcomes, Learning Techniques.