Abstract

The design and development of a web application to identify a high or low probability of student dropout at the National Learning Service (SENA) in Colombia, aiming to streamline the process of identifying and supporting potential candidates for assistance provided by the institution through the student welfare department. Throughout the development, socioeconomic variables with the highest impact on characterized academic dropout processes to create a dataset. This dataset was then utilized with various artificial intelligence techniques explored in Machine Learning (Decision Trees, K-means, and Regression), ultimately determining the most effective algorithm for integration into the Software. The decision tree classification technique emerged as the most effective, achieving an impressive accuracy of 91% and a minimal error rate of 9%, substantiating its state-of-the-art standing. As a result, this Software has optimized processes within the Student Welfare Department at SENA and is adaptable for use in any higher education institution.