Aplicação de clustering para apoio à análise de risco de desequilíbrio econômico-financeiro nas atas de registro de preço da fufmt

Abstract

This study investigates the economic-financial rebalancing process in Price Registration Agreements within the Federal University of Mato Grosso (UFMT), proposing the application of statistical analysis and machine learning techniques to optimize decision-making and the management of public procurement. The research develops a technological tool, implemented in Python, capable of classifying and clustering the institution’s bidding documents, such as public notices, agreements, and supply orders, based on their textual similarity, aiming to identify patterns related to the occurrence of economic-financial imbalance in contracts. The study adopts a quantitative approach through modeling and simulation, as well as the implementation of clustering algorithms, with emphasis on K-means, enabling the analysis of unstructured data. The results demonstrate the feasibility of applying artificial intelligence (AI) techniques as a decision-support tool in public management, allowing the identification of contracts with a higher likelihood of instability and providing a foundation for preventive monitoring of procurement processes. The research proposal contributes to improving administrative efficiency, transparency, and sustainability in the use of public resources, offering a framework that can be replicated in other public institutions.