Resumen

TITO, Anthony Edwin Aco; CONDORI, Bryan Orlando Hancco  y  VERA, Yasiel Pérez. Comparative analysis of Machine Learning Techniques for the prediction of cases of university dropout. RISTI [online]. 2023, n.51, pp.84-98.  Epub 30-Sep-2023. ISSN 1646-9895.  https://doi.org/10.17013/risti.51.84-98.

University dropout has a detrimental impact on numerous students; this phenomenon may be associated with personal issues, economic constraints, and other factors. Given this situation, the importance of developing a predictive model for such cases arises. To achieve this, Machine Learning techniques were proposed and employed, including Logistic Regression, Naive Bayes, Multilayer Perceptron Neural Network, Decision Tree, Support Vector Machine, and Random Forest. A dataset was selected and underwent data cleaning, addressing missing values and outliers. Subsequently, records with the 'Enrolled' outcome variable were removed, focusing solely on 'Dropout' and 'Graduate' categories. Each model was trained and tested using cross-validation with folds. Ultimately, they were compared based on accuracy, precision, and recall metrics, leading to the conclusion that Logistic Regression is the technique that yields the best results for predicting university dropout in the considered dataset.

Palabras clave : Comparative analysis; University dropout; Logistic Regression; Machine Learning; Prediction.

        · resumen en Español     · texto en Español     · Español ( pdf )