Systematic Literature Review on Predictive Modeling with a Governmental Focus

Abstract

This systematic review investigates the use of predictive models in the governmental context, focusing on their applications, methods, implementation areas, and reported results. Five guiding research questions were addressed: the most commonly used predictive models, the domains of public administration where they are applied, the most impacted areas, the functions performed by AI, and the gaps in the current literature. The search was conducted in the Scopus and Web of Science databases (2020 to 2025), resulting in 1,020 studies, of which 45 met the inclusion, exclusion, and quality criteria. The analysis revealed the frequent use of neural networks, decision trees, random forests, and natural language processing, applied in areas such as health, education, finance, urban planning, and e-governance. In summary, the application of predictive models in the governmental context offers great potential, but also exposes significant challenges that require continuous strengthening of infrastructure and governance. The integration of these models must be carefully monitored to ensure equity and transparency in the use of AI technologies in the public sector.