Predictive Identification of Key Factors for Student Retention Using Machine Learning Techniques to Strengthen Academic Support and Educational Success

The FIRETA project aims to identify the critical variables influencing student dropout rates at the Universidad Politécnica Salesiana del Ecuador, across both face-to-face and online modalities. Facing a dropout rate of approximately 10%, particularly in early academic levels, the research proposes a data science-driven approach to transform institutional academic management. The methodology integrates the analysis of socioeconomic records, academic histories, and demographic data using Machine Learning algorithms, including k-NN and neural networks, alongside structural equation modeling. The study seeks not only to predict dropout risk but also to segment students to implement personalized retention strategies. As a result, the project expects to develop predictive tools that allow the university to optimize resources, improve graduation rates, and strengthen its academic reputation. The project aligns with the SDG for Quality Education, promoting a culture of evidence-based decision-making and continuous improvement.<br/><br/><b>Goal</b>: <br/>Determine the factors influencing student retention at the Universidad Politécnica Salesiana del Ecuador using supervised and unsupervised Machine Learning techniques. The goal is to strengthen academic support and improve educational success.<br/><br/><b>Research lines</b>: <br/>Technologies applied to education