Prediction and Learning Analysis Using Ensemble Classifier Based on GA in SPOC Experiments

The teaching mode combining Massive Open Online Course (MOOC) with flipped classroom has been emerged in recent years since the arrangement can enhance obviously students’ learning outcome. In this paper, we proposed an ensemble approach based on genetic algorithm (GA) for feature selection (EA-GA) for MOOC data analysis, focusing on the prediction of students’ learning outcome. The work is based on the implementation of an online course from a college. The tracking data is collected from both the online MOOC platform and the offline classroom. After combining all data together, a GA based ensemble system is designed to predict students’ academic performances. Some other machining learning algorithms are also derived for performance comparison of different algorithms. Simulation results showed the proposed the EA-GA preforms better than other algorithms to predict well the students’ learning score. The “shared features” found by EA-GA from massive features are helpful to discriminate at-risk students and excellent students for different teaching intervention purpose.