Resumen

PEREZ-MONTERO, Eilen-Lorena  y  QUIMBAYO-CASTRO, Julián-Andrés. Prediction of Academic Achievement in Mirror Classes: Sociodemographic and Pedagogical Analysis through Data Mining. Rev. Cient. [online]. 2024, n.49, pp.79-98.  Epub 19-Jun-2024. ISSN 0124-2253.  https://doi.org/10.14483/23448350.21820.

The aim of this study was to generate predictive models of academic achievement in mirror classes at different levels using the sociodemographic and pedagogical profile of 103 students and data mining algorithms. A quantitative approach was followed based on the CRISP-DM (cross-industry standard process for data mining) model, addressing phases such as problem definition; data acquisition, understanding, and analysis; feature extraction; modeling; model evaluation; and deployment. The results show that the best model corresponds to the support vector classifier (SVC) algorithm, with an accuracy of 62%, greater robustness to changes, and 0.89 accuracy when classifying achievement, which represents a reduction of 0.30 compared to the multinomial logistic regression model (RLOG) and the linear discriminant analysis (LDA) algorithm. The F1 score metric has a balance of 0.64. Important predictive attributes were identified to determine academic achievement with respect to the demographic variable, such as age, gender, semester, and whether the student lives with their parents. Likewise, in the pedagogical variable, aspects such as stimuli to participation, punctuality, collaborative work, clarity of activities and feedback, ease of expression of the teacher, resources to support learning, and technological means of communication were highlighted. This study proposes computer tools for universities to design improvement and prevention strategies regarding academic performance, based on the implementation of the internationalization strategy called mirror class in systems engineering students of a private university in Huila (Colombia).

Palabras clave : academic achievement; data mining; higher education; mirror class; predictive model.

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