https://repositorio.ifg.edu.br/handle/prefix/61 Sat, 12 Sep 2026 09:51:26 GMT 2026-09-12T09:51:26Z https://repositorio.ifg.edu.br/handle/prefix/2550 Título: Avaliação de Modelos Generativos para Extração de Informações em Processos da Pró-Reitoria de Desenvolvimento Institucional e Recursos Humanos do Instituto Federal de Goiás Autor(es): Silva, Daniel Vitor Ferreira Primeiro Orientador: Sousa, Daniel Xavier de Abstract: Efficiency in administrative management represents a growing challenge in the public sector, particularly given the volume of unstructured documents requiring manual analysis. At the Federal Institute of Goiás (IFG), processes such as Training Leave require staff to dedicate significant time locating data scattered across attachments and dispatches. This work proposes and evaluates a solution based on Large Language Models (LLMs) for the automatic Information Extraction (IE) from these proceedings. The methodology adopted a two-stage architecture (Generation and Validation) and investigated the impact of four text selection strategies (Full Body, Requirement-only, Sentence Chunking, and Token Chunking) on the performance of both commercial (GPT family and DeepSeek) and open-source models (GPT-OSS, DeepSeek). Experiments demonstrated that the text selection strategy exerts a greater influence on extraction quality than the model choice itself. The "Full Body" approach proved to be the most robust, allowing models to capture complex semantic relationships that are lost in chunking strategies. Results indicate that while commercial models like GPT-4o-mini offer greater overall consistency, free models, when paired with a broad context strategy, achieve competitive similarity indices (exceeding 0.90 in specific scenarios). This study delivers a validated analysis systematic for this domain, demonstrating the technical feasibility of intelligent automation in personnel management. Editor: Insitituto Federal de Educação, Ciência e Tecnologia de Goiás Tipo: Trabalho de Conclusão de Curso Wed, 10 Sep 2025 00:00:00 GMT https://repositorio.ifg.edu.br/handle/prefix/2550 2025-09-10T00:00:00Z https://repositorio.ifg.edu.br/handle/prefix/2425 Título: Comparação entre abordagem neurais e não neurais para modelagem de tópicos em tweets sobre Institutos Federais Brasileiros. Autor(es): Dias, Paulo David Silva Primeiro Orientador: Canuto, Sérgio Daniel Carvalho Abstract: The analysis of public perception of educational institutions on social media can reveal topics of community interest and support improvements in communication and academic management. Topic Modeling (TM) techniques allow for the automation of this identification, but studies evaluating aspects such as coherence, diversity, stability, and computational time of TM-generated topics in the educational context remain scarce. This work investigates these dimensions by applying TM to a novel corpus of tweets about Brazilian Federal Institutes, segmented by five macro-regions. The results show that, although traditional methods such as LDA and NMF perform better in terms of diversity and stability, respectively, the BERTopic approach with the BERTweet-BR model produced topics with higher coherence compared to LDA. Editor: Insitituto Federal de Educação, Ciência e Tecnologia de Goiás Tipo: Trabalho de Conclusão de Curso Thu, 27 Mar 2025 00:00:00 GMT https://repositorio.ifg.edu.br/handle/prefix/2425 2025-03-27T00:00:00Z https://repositorio.ifg.edu.br/handle/prefix/2286 Título: Tecnologia Assistiva para Bengalas de Pessoas com Deficiência Visual Autor(es): Oliveira, Rafael Souza Primeiro Orientador: José, Alexandre Bellezi Abstract: This study investigates assistive technologies that support visually impaired individuals, with a particular focus on mobility. Assistive technologies play a crucial role in reducing mobility barriers, promoting greater autonomy and safety. Despite the existence of various solutions, a preliminary study revealed that the traditional white cane remains the primary mobility tool used, highlighting a gap in more advanced options to enhance user perception. To address this gap, this study analyzes the mobility challenges and existing tools. Based on this analysis, a low-cost solution is proposed: an enhanced cane integrating sensors, actuators, and an ESP32 microcontroller. This proposal aims to offer an alternative to improve the locomotion of visually impaired people. To validate this approach, a prototype of the cane was implemented. Editor: Insitituto Federal de Educação, Ciência e Tecnologia de Goiás Tipo: Trabalho de Conclusão de Curso Tue, 25 Mar 2025 00:00:00 GMT https://repositorio.ifg.edu.br/handle/prefix/2286 2025-03-25T00:00:00Z https://repositorio.ifg.edu.br/handle/prefix/2282 Título: Análise biobjetivo de roteamento de fluxo em redes com custos dos fluxos e qualidades dos enlaces variados Autor(es): Nascimento, Gabriel Henrique do Primeiro Orientador: Fernandes, Kátia Cilene Costa Abstract: This work begins with a study on multiobjective optimization with emphasis on biobjective problems and methods to find the solution of these problems, with emphasis on the method called ε-constraint. The proposal is to study the algorithm to solve a biobjective network flow routing problem, presented in (PINTO; FERNANDES; CARDOSO, 2021), considering the costs of varied links. This algorithm generates a minimal complete set of Pareto-Optimal solutions. The metrics evaluated are the cardinality of this minimum complete set, number of iterations and the execution time of this algorithm. All these metrics will be compared both for the fixed cost scenario presented in the work (PINTO; FERNANDES; CARDOSO, 2021) and for the varied costs implemented in this present work. The network model adopted for routing flows is based on the Barabási-Albert model. Editor: Insitituto Federal de Educação, Ciência e Tecnologia de Goiás Tipo: Trabalho de Conclusão de Curso Mon, 19 Dec 2022 00:00:00 GMT https://repositorio.ifg.edu.br/handle/prefix/2282 2022-12-19T00:00:00Z