A Comparative Study of Artificial Intelligence, Machine Learning, and Deep Learning Approaches in Predicting Academic Performance

Abstract: Academic performance prediction has become a crucial area in educational data mining, enabling early intervention for at-risk students and improving institutional strategies. This study provides a comparative analysis of Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) approaches in predicting academic performance by exploring a range of models, including linear regression, random forests, support vector machines, multilayer perceptrons, and convolutional neural networks. The analysis leverages demographic, socioeconomic, and historical academic data to evaluate model effectiveness using metrics such as accuracy, mean absolute error, and root mean squared error. While ML algorithms like Support Vector Regressors offer robust predictive results, the incorporation of deep learning models (especially CNNs and Bi-LSTM), demonstrates improved performance -- albeit with increased computational complexity and data requirements. The review identifies the suitability of each method depending on the context, scale, and features of the educational dataset. The findings suggest that deep learning models excel in handling complex, high-dimensional data, whereas traditional ML models perform reliably with structured and tabular information. Ultimately, integrating comprehensive preprocessing and feature engineering enhances results, signalling the need for tailored approaches in educational settings.

Keywords: Academic Performance Prediction, Machine Learning, Deep Learning, Artificial Intelligence, Educational Data Mining