The use of machine learning techniques has transformed various sectors, including education, by allowing for more accurate prediction of students’ academic preferences. The main objective of this study is to compare the effectiveness of classical algorithms, such as Logistic Regression, Support Vector Machines (SVM), and K-Nearest Neighbors (KNN), with advanced algorithms, such as Random Forest and XGBoost, in predicting academic program choices in students at a technical university in Ecuador. The methodology used included the use of a dataset of 1,170 student responses, which were processed and analyzed using these algorithms. The techniques were evaluated using metrics such as precision, recall, and F1-score. The results obtained indicate that the advanced algorithms, particularly XGBoost, significantly outperform the classical ones in terms of precision, reaching 89.3%. Although the classical algorithms demonstrated faster execution times, their lower precision makes them less suitable for complex prediction tasks. In conclusion, advanced algorithms are presented as more effective tools for academic planning and optimization of educational resources, providing the possibility of personalizing learning and improving decision-making in educational institutions.
