Predicting high school students’ performance on e-learning using experience application interface (xapi) and artificial intelligence

Abstract

Motivated by the importance of education to the development of individuals and society, this research investigates using Artificial Intelligence (AI) to predict student performance. This thesis pays special attention to the current problem by evaluating the performance once AI is used as tool to add value to existing teaching strategies for long-term improvements. Specifically, the thesis explores applications of Artificial Intelligence in education: predicting student performance. The thesis identifies potential challenges, current limitations, and suggestions for further improvement. Rigorous analysis using Artificial Intelligence will lay a solid foundation for further study within the domain. This research discusses the possibility of using xAPI (experience Application Interface) for information collection for Artificial Intelligence dataset input to predict student performance with an e-learning management system. This dissertation pays special attention to the factors influencing academic performance and, viable ways to anticipate and predict students’ long-term achievements. This research provides a way of anticipating student’s results using a machine learning-based system, to predict students’ marks. Using Jupyter Notebook from Anaconda software the dataset is processed. Using 4 types of models namely Logistic Regression (LR), Support Vector Regression (SVR), K Nearest Neighbours (KNN), and Artificial Neural Network (ANN), the results predicted are in the range between 40 and 85 out of 100, with an accuracy of 95%. From the results also we detected the need of increasing the dataset size, especially increasing quantitative dependent variables for prediction. During research we also realized that e-learning can’t be implemented alone, hence the requirement of a blended teaching system.