• JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
  • JoomlaWorks Simple Image Rotator
 
  Bookmark and Share
 
 
Master's Dissertation
DOI
https://doi.org/10.11606/D.3.2021.tde-14022022-122906
Document
Author
Full name
Rodrigo Amorim Ruiz
E-mail
Institute/School/College
Knowledge Area
Date of Defense
Published
São Paulo, 2021
Supervisor
Committee
Bona, Glauber De (President)
Finger, Marcelo
Soares, Anderson da Silva
Title in English
Jurisprudence search based on facts similarity using NLP and ML techniques.
Keywords in English
Artificial intelligence
Bag-of words
Cosine similarity
Deep learning
FastText
GloVe
Jurisprudence
Logistic regression
Long short-term memory
Machine learning
Multilayer perceptron
Naive bayes
Natural language processing
Neural network
Siamese neural network
TF-IDF
Transfer learning
Transformer
Word embedding
Word2Vec
Abstract in English
Part of a lawyers job is to understand the clients problem, to textually describe its facts and to apply the sources of law. To support a new legal case, a handful of past judgments on similar cases are typically employed by the lawyers, but finding them is currently a time-consuming procedure. To address this problem, we built a machine learning model responsible for classifying similarity between two facts descriptions. This similarity metric measures how much (from 0 to 1) a legal decision may be used to support another. We trained different model architectures combining several state-of-the-art natural language processing and machine learning techniques using an extracted dataset from the Superior Court of Justice website of past judgments, which enabled the dynamic construction of facts description pairs when one case cites another as a reference. The final best architecture employs TF-IDF for encoding and reducing dimensionality of our input documents, a Siamese Neural Network (SNN) with a Multilayer Perceptron (MLP) for feature extraction and a final layer, another MLP, responsible for concatenating and classifying the features into the similarity metric, achieving 85.98% accuracy, 83.89% precision and 89.06% recall. Such a model would enable the lawyer to compare a case facts description with several judgments of the jurisprudence and start their search on the most similar ones.
Title in Portuguese
Pesquisa de jurisprudência baseada na semelhança de fatos usando técnicas de PNL e ML.
Keywords in Portuguese
Aprendizado computacional
Inteligência artificial
Jurisprudência
Linguagem Natural
Redes neurais
Abstract in Portuguese
Parte do trabalho de um advogado é entender o problema do cliente, descrever textualmente seus fatos e aplicar as fontes da lei. Para apoiar um novo processo legal, uma série de julgamentos anteriores em casos semelhantes são normalmente empregados pelos advogados, mas encontrá-los é atualmente um procedimento que demanda tempo. Para resolver esse problema, construímos um modelo de aprendizado de máquina responsável por classificar a similaridade entre as descrições de dois fatos. Essa métrica de similaridade mede quanto (de 0 a 1) uma decisão legal pode ser usada para apoiar outra. Treinamos diferentes arquiteturas combinando várias técnicas de processamento de linguagem natural e aprendizado de máquina do estado da arte usando um conjunto de dados extraído do site do Superior Tribunal de Justiça de julgamentos anteriores, o que possibilitou a construção dinâmica de pares de descrição de fatos quando um caso cita outro como referência. A melhor arquitetura final emprega TF-IDF para codificar e reduzir a dimensionalidade dos documentos de entrada, uma Rede Neural Siamesa (SNN) com um Multilayer Perceptron (MLP) para extração de "features" e uma camada final, outro MLP, responsável por concatenar e classificar essas "features" na métrica de similaridade, alcançando 85,98% de acurácia, 83,89% de precisão e 89,06% de sensibilidade. Tal modelo permitiria ao advogado comparar a descrição dos fatos de um caso com vários julgamentos da jurisprudência e iniciar sua busca pelos mais semelhantes.
 
WARNING - Viewing this document is conditioned on your acceptance of the following terms of use:
This document is only for private use for research and teaching activities. Reproduction for commercial use is forbidden. This rights cover the whole data about this document as well as its contents. Any uses or copies of this document in whole or in part must include the author's name.
Publishing Date
2022-02-21
 
WARNING: Learn what derived works are clicking here.
All rights of the thesis/dissertation are from the authors
CeTI-SC/STI
Digital Library of Theses and Dissertations of USP. Copyright © 2001-2024. All rights reserved.