Evaluating Machine Learning Techniques for Detecting Learning Difficulties: A Focus on Moroccan Baccalaureate Students

In recent years, researchers have begun exploring how machine-learning techniques can be applied to education, particularly in the identification of students with learning difficulties. This article aims to determine the most accurate and precise machine learning technique for identifying learning difficulties among second year baccalaureate students way before the final examination. Consequently, three academic courses are selected, each forming a separate dataset, and three classification algorithms are tested. The selected models aims to determine the necessity of supplementary courses for students at risk, by offering tailored recommendations on whether and when additional courses may be beneficial. The study's findings showed that, in comparison with other techniques, Decision Tree algorithm achieved the best outcomes.