Predictive model based on artificial intelligence and contextual awareness to identify students at risk of dropping out of university in Panama

Introduction: This article presents the findings from the research on the “Analysis, design, and development of a context-aware intelligent system for university dropout prediction using a microservices architecture” conducted at the Technological University of Panama in 2024. The study focuses on creating a predictive model using artificial intelligence that integrates students’ academic, sociodemographic, and psychological context to identify students at risk of early dropout. Problem: In Latin America, the educational system struggles with high dropout rates, particularly at the university level. The early years of university education are crucial, and dropping out during this period significantly impacts a country’s social, labor, and economic development. Objective: The main goal of this research was to develop a context-aware predictive model using artificial intelligence to identify students at risk of dropping out at the Technological University of Panama. The model incorporates academic, socioeconomic, and psychological factors to predict dropout risks more accurately. Methodology: The study follows the CRISP-DM (Cross-Industry Standard Process for Data Mining) methodology, starting with problem understanding and followed by the collection and processing of students’ academic, socioeconomic, and psychological data. Three machine learning models were implemented and evaluated: Logistic Regression, Random Forest, and Gradient Boosting. Each model was tested under different scenarios to identify the most effective one. Results: The Random Forest model in Scenario 3 demonstrated the best performance, offering a strong balance between accuracy and generalization. This model was the most effective for predicting university dropouts based on the collected data. Conclusion: The developed predictive model is a relevant and innovative tool that can help the Technological University of Panama identify students at risk of dropping out early. By using artificial intelligence and a context-aware approach, this tool can contribute to reducing dropout rates and improve student retention. Originality: This research introduces a novel approach to university dropout prediction by integrating context-awareness with artificial intelligence. The multidimensional nature of the model, considering academic, sociodemographic, and psychological factors, sets it apart from traditional predictive models. Limitations: One of the key limitations of this research was the restricted availability of real psychological data, which limited the comprehensiveness of the model.