Menezes Neto, Elias Jacob deBrito, Bruna Alice Oliveira de2023-11-222023-11-222023-10-26BRITO, Bruna Alice Oliveira de. Utilização de técnicas de processamento de linguagem natural para identificação automática de doenças em processos da JFRN. 2023. 59 f. Trabalho de Conclusão de Curso (Especialização em Residência em Tecnologia da Informação) - Instituto Metrópole Digital, Universidade Federal do Rio Grande do Norte, Natal, 2023.https://repositorio.ufrn.br/handle/123456789/55408The jurisdiction exercised by the Federal Court of Rio Grande do Norte (JFRN) is extensive and encompasses the analysis and resolution of a wide range of judicial cases of various natures, which may pertain to the environment, social security, tax law, among others [1]. In these cases, on one side, there are private individuals, and on the other side, there are the Union, public enterprises, federal autarchies and foundations, or professional oversight councils [1]. In this context, the integration of Artificial Intelligence (AI) is of significant interest in the field. One of the crucial tools in this scenario is Natural Language Processing (NLP), as numerous legal procedures involve the analysis and interpretation of textual documents. Named Entity Recognition (NER) is one area of NLP dedicated to recognizing and classifying entities mentioned in texts. NER gains special prominence in the legal field, where a multitude of legal documents, contracts, petitions, jurisprudence, and other textual genres require meticulous analysis. Given the foregoing, the main objective of this work is the application of NLP techniques in the processes of the Special Federal Courts of the JFRN, and the secondary objective is the construction of BI dashboards for visualizing the data applied in these techniques. This application analyzed the texts of initial petitions and judgments, and two distinct models were developed. The first model aims to discern whether a process is related to healthcare or not, using an XGBoost model. The second model is responsible for identifying and highlighting words and terms denoting some form of ailment, using a pre-trained BERT (Bidirectional Encoder Representations from Transformers) model, specifically BioBERTpt for disease recognition [3]. By implementing these models, the goal was to facilitate the identification of healthcare-related issues within the texts so that responsible officials could have an overview of the diseases found and optimize the allocation of resources.NERPLNProcessos de saúde.bachelorThesisCNPQ::CIENCIAS EXATAS E DA TERRA::CIENCIA DA COMPUTACAO::METODOLOGIA E TECNICAS DA COMPUTACAO