Text Classification in Legal Documents Extracted from Lawsuits in Brazilian Courts

• André Aguiar UNIFOR https://orcid.org/0000-0001-5437-0945
• Raquel Silveira IFCE https://orcid.org/0000-0001-7445-605X
• Vládia Pinheiro UNIFOR https://orcid.org/0000-0002-9851-8304
• Vasco Furtado UNIFOR https://orcid.org/0000-0001-8721-4308
• João Araújo Neto UNIFOR https://orcid.org/0000-0002-0690-2449

Resumo

Recently, Brazil’s National Council of Justice (CNJ) highlighted the importance of robust solutions to perform automated lawsuit classification. A correct lawsuit classification substantially improves the assertiveness of (i) distribution, (ii) organization of the agenda of court hearing and sessions, (iii) classification of urgent measures and evidence, (iv) identification of prescription and (v) prevention. This paper investigates different text classification methods and different combinations of embeddings, extracted from Portuguese language models, and information about legislation cited in the initial documents. The models were trained with a Golden Collection of 16 thousand initial petitions and indictments from the Court of Justice of the State of Ceará, in Brazil, whose lawsuits were classified in the five more representative CNJ’s classes - Common Civil Procedure, Execution of Extrajudicial Title, Criminal Action - Ordinary Procedure, Special Civil Court Procedure, and Tax Enforcement. Our best result was obtained by the BERT model, achieving 0.88 of F1 score (macro), in the experiment scenario that represents the lawsuit in an embedding formed by concatenating the texts of all the petitions that contain at least one citation to one legislation. Legal documents have specific characteristics such as long documents, specialized vocabulary, formal syntax, semantics based on a broad specific domain of knowledge, and citations to laws. Our interpretation is that the representation of the document through contextual embeddings generated by BERT, as well as the architecture of the model with bidirectional contexts, makes it possible to capture the specific context of the domain of legal documents.

Palavras-chave: Text classification, Legal domain, Brazilian lawsuits, Language models