Student At-Risk Identification and Classification Through Multitask Learning: A Case Study on the Moroccan Education System

Early identification of students at-risk of dropping out or failing is a critical challenge in education. In this paper, we address this imperative by deploying multitasking models with a dual purpose: early identification and subsequent classification into success, failure, or dropout. The multitask models used include the deep multitask model, as well as and based multitask models. We provided a comparative analysis of the performance of traditional machine learning, deep learning, and multitasking models. Our proposed approach has been applied to the Moroccan education system using a proprietary dataset provided by the Moroccan Ministry of National Education, Pre-school, and Sports. The results show that the multitasking model demonstrates superior accuracy and effectiveness in both tasks, while also reducing training and prediction time.