Banca de QUALIFICAÇÃO: Wagner Miranda Costa

Uma banca de QUALIFICAÇÃO de MESTRADO foi cadastrada pelo programa.
STUDENT : Wagner Miranda Costa
DATE: 28/01/2023
TIME: 09:00
LOCAL: remoto
TITLE:

Support in the Formulation of TCU Jurisprudence Using Bag-of-Concepts for the Analysis of Semantic Similarity Judgments

KEY WORDS:

Natural Language Processing, Information Retrieval, Document Vector Representation, Bag-of-Concepts, Word Embeddings

PAGES: 45
BIG AREA: Ciências Exatas e da Terra
AREA: Ciência da Computação
SUBÁREA: Metodologia e Técnicas da Computação
SUMMARY:

Jurisprudence refers to the set of decisions and interpretations of law made by courts regarding a particular topic. At the Federal Court of Accounts (TCU) the elaboration of jurisprudence is defined from a grouping of similar judgments. Therefore, it is very important to develop automated tools that help specialists in the search for similar judgments. However, this is a challenging task for the computing area due to the peculiar language present in the texts of the judgments and the massive volume of data. Therefore, this work aims to present a study and investigation of the Bag-of-Concepts (BoC) approach for the representation of these textual documents. The BoC approach generates a dictionary of concepts by grouping word-vectors generated by some word-embedding technique, and calculates the frequencies of these concepts to represent the documents in vectors. The distance between these vectors, then, is used as a criterion to analyze the similarity between the documents. The purpose of this work is to investigate extended solutions for the BoC technique that contemplate the peculiar characteristics of the TCU judgments, in order to make the document vectors more discriminative and, thus, return more documents closer to the user's expectation.

BANKING MEMBERS:
Presidente - 3064724 - GLAUCO VITOR PEDROSA
Interno - 1937247 - BRUNO CESAR RIBAS
Interno - 1821656 - THIAGO DE PAULO FALEIROS
Externo à Instituição - EDUARDO DE PAULA COSTA - USP