Citas bibligráficas
Borjas, N., Saqui, C. (2023). Modelo de machine learning para disminuir la tasa de deserción de estudiantes antiguos en centro de estudio superior [Trabajo de Suficiencia Profesional, Universidad Peruana de Ciencias Aplicadas (UPC)]. http://hdl.handle.net/10757/672020
Borjas, N., Saqui, C. Modelo de machine learning para disminuir la tasa de deserción de estudiantes antiguos en centro de estudio superior [Trabajo de Suficiencia Profesional]. PE: Universidad Peruana de Ciencias Aplicadas (UPC); 2023. http://hdl.handle.net/10757/672020
@misc{renati/413821,
title = "Modelo de machine learning para disminuir la tasa de deserción de estudiantes antiguos en centro de estudio superior",
author = "Saqui Marin, Carlos Jean Pierre",
publisher = "Universidad Peruana de Ciencias Aplicadas (UPC)",
year = "2023"
}
The project aims to implement a machine learning model in higher education institutions. The main focus is on optimizing a process in the student experience area, specifically, predicting the student dropout rate. To develop the project, a comparison of machine learning algorithms was conducted with the purpose of selecting the most suitable one. The model design included creating a master table that incorporates all variables such as current academic information, past academic information, grade information, attendance ratios, personal information, and survey information, along with statistical analysis using techniques like WOE to identify which of these variables significantly contribute to a prediction. Validations were performed by comparing algorithms like RUSBost, GausNB, Percept, LightGBM, using key indicators like Recall mean, Precision mean, ROC AUC mean, and other indicators to ensure the model is effective and accurate. Therefore, after evaluating all the algorithms, LightGBM obtained the highest score with a Recall mean of 86.37%, Precision mean of 86.83%, and ROC AUC mean of 77.61%. Additionally, a continuity plan has been established to improve the machine learning model over time. Before implementation, the student dropout rate was 13.45% in the 2022-1 period, and after implementation in the 2023-1 period, it decreased to 9.45%.
Este ítem está sujeto a una licencia Creative Commons Licencia Creative Commons