Red neuronal para predecir el rendimiento académico

Abstract

This study presents a model designed to predict academic performance using neural networks. It is framed within a quantitative approach and is categorized as a multivariate correlational study. The research is based on a database from an educational institution, available in the data repository of the University of California, Irvine. R was chosen as the programming language, with RStudio as the development environment. The CRISP-DM methodology was adopted to carry out the data analysis. The construction of the neural network was carried out using the nnet package, available in the Comprehensive R Archive Network (CRAN). The neural network model was applied to data collected from 649 students, and its predictive ability was comprehensively evaluated. After comparing it with a multiple linear regression model, it was observed that the neural network model achieved an effectiveness of 87% in predicting academic performance, evidencing its suitability for this purpose.