Análisis de factores de la deserción de estudiantes SENA en los niveles técnico y tecnólogo, mediante metodologías de Machine Learning
No hay miniatura disponible
Fecha
2023
Director de trabajo de grado
Título de la revista
ISSN de la revista
Título del volumen
Editor
Universidad de La Salle. Escuela de Negocios. Maestría en Analítica e Inteligencia de Negocios
Descripción
El Servicio Nacional de Aprendizaje SENA ha graduado a millones de colombianos altamente calificados para contribuir al crecimiento laboral e industrial del país. No obstante, la institución no ha sido ajena al fenómeno de deserción estudiantil, ya que en el año 2022 reportó más del 25% de deserción de sus estudiantes inscritos, es decir, que uno de cada cuatro estudiantes no consigue obtener su titulación debido a factores socioeconómicos, familiares, laborales y /o geográficos. En el presente trabajo de investigación, se han utilizado técnicas analíticas basadas en metodologías de Machine Learning para identificar los factores que influyen en la decisión de un aprendiz de abandonar su proceso formativo. El objetivo es aumentar la tasa de retención y brindar ayuda a más colombianos para que completen sus estudios. Con el fin de obtener dichos resultados, se implementaron varios modelos para identificar los factores que influyen en la decisión de un estudiante de abandonar sus estudios. En total, se desarrollaron cinco modelos diferentes para predecir y caracterizar a los aprendices. De todos ellos, el modelo de Random Forest demostró un rendimiento destacado, con un nivel de precisión del 90,27%. Este modelo permitió identificar los factores principales que influyen en la caracterización de un estudiante y que pueden llevarlo a tomar la decisión de desertar.
The National Apprenticeship Service (SENA) has graduated millions of highly skilled Colombians to contribute to the country's labor and industrial growth. However, the institution has not been immune to the phenomenon of student dropout, as in 2022 it reported a dropout rate of over 25% among its enrolled students, meaning that one in every four students fails to obtain their degree due to socioeconomic, familial, work-related, and/or geographical factors. In this research work, analytical techniques based on Machine Learning methodologies have been used to identify the factors that influence an apprentice's decision to abandon their educational process. The objective is to increase the retention rate and provide assistance to more Colombians in completing their studies. To achieve these results, several models were implemented to identify the factors influencing a student's decision to drop out. In total, five different models were developed to predict and characterize the apprentices. Among them, the Random Forest model demonstrated outstanding performance, with an accuracy level of 90.27%. This model allowed for the identification of the main factors that influence the characterization of a student and can lead them to make the decision to drop out.
The National Apprenticeship Service (SENA) has graduated millions of highly skilled Colombians to contribute to the country's labor and industrial growth. However, the institution has not been immune to the phenomenon of student dropout, as in 2022 it reported a dropout rate of over 25% among its enrolled students, meaning that one in every four students fails to obtain their degree due to socioeconomic, familial, work-related, and/or geographical factors. In this research work, analytical techniques based on Machine Learning methodologies have been used to identify the factors that influence an apprentice's decision to abandon their educational process. The objective is to increase the retention rate and provide assistance to more Colombians in completing their studies. To achieve these results, several models were implemented to identify the factors influencing a student's decision to drop out. In total, five different models were developed to predict and characterize the apprentices. Among them, the Random Forest model demonstrated outstanding performance, with an accuracy level of 90.27%. This model allowed for the identification of the main factors that influence the characterization of a student and can lead them to make the decision to drop out.
Palabras clave
Aprendizaje combinado, Deserción escolar, Aprendizaje automático, Análisis de tendencia, Caracterización, Tecnología educacional