Virtual Booth

Educational Data Mining (EDM) and Data Science are the emerging technologies in the education sector to deal with student-related data and extract useful academic performance information. With the advent of Deep Learning (DL) technology, predicting graduation rates in terms of course grades, Grade Point Average (GPA), and risks is essential and demanding in the MOOC platform. In this project, the MOOC platform has been developed with frontend and backend using different technologies. The prototype has been trained using real dataset collected from the student academic data logs. The custom-built model has been trained, validated, and tested using student data logs (1670 records). The proposed hybrid Convolutional Neural Networks + Long Short-Term Memory (CNN+LSTM) has been developed. The system contributed to identify the features and target classes for predictive learning analytics with respect to students’ examination data. The eXplainable Artificial Intelligence (XAI) algorithms, namely SHAP (SHapley Additive exPlanations) were integrated with the proposed hybrid DL model to determine significant factors. This system provides comprehensive detail of student academic and course management data with frontend (React Native) and backend (Mongo DB Database) for student, instructor and admin users. Each student can check his/or her predicted CGPA and risk assessment using Single Prediction page. This system successfully predicted the student's final cumulative GPA and recommendations using previous academic records with highest accuracy of 92.2%. This system provides support for the education stakeholders during early-warning and risk-assessments for students’ academic development in the MOOC platform through visualized graphs and statistics for instructor and admin users in their respective console panels.