Banca de DEFESA: Ilo César Duarte Cabral

Uma banca de DEFESA de MESTRADO foi cadastrada pelo programa.
STUDENT : Ilo César Duarte Cabral
DATE: 26/02/2024
TIME: 09:00
LOCAL: remoto
TITLE:

Mining of Federal Legislative Data for Analysis and Prediction of Lawmaking


KEY WORDS:
data mining, legislative data, machine learning, natural language processing, imbalanced dataset
 
 

PAGES: 90
BIG AREA: Ciências Exatas e da Terra
AREA: Ciência da Computação
SUBÁREA: Metodologia e Técnicas da Computação
SUMMARY:

The release of legislative data by the Brazilian government opened an opportunity to understand aspects related to the legislative process. By analyzing historical patterns and relevant variables, it is possible to anticipate legislative results, optimizing the decision-making process. Predicting the votes of deliberative bodies, for example, can lead to a better understanding of government policies and thus generate actionable strategies, allowing legislators to identify critical issues, allocate resources efficiently and anticipate possible impasses. This work set out to investigate models for analysis and prediction that maximize the use of publicly accessible heterogeneous data from legislative data to understand the approval/disapproval of legislative proposals. To this end, classification models based on machine learning algorithms and natural language processing were developed on categorical, textual and processing data of Legislative Proposals, in order to identify discriminatory factors that could influence the approval of Bills and Amendment Projects. As a contribution, the classification models were evaluated in five scenarios, using different sets of attributes. The results obtained show an F1-Score of up to 70\% considering only the categorical data of the propositions and, when aggregating the processing data, it is possible to obtain an F1-Score of up to 97\%. The tests carried out demonstrate the feasibility of predicting the approval of a proposition during its flow in the legislative process, generating results that add knowledge and lead to a better understanding of aspects related to the Brazilian legislative process at the federal level.


COMMITTEE MEMBERS:
Presidente - 3064724 - GLAUCO VITOR PEDROSA
Interno - 3089262 - JOHN LENON CARDOSO GARDENGHI
Interno - 3128249 - LUIS PAULO FAINA GARCIA
Externo à Instituição - EDUARDO DE PAULA COSTA - USP
Notícia cadastrada em: 06/02/2024 18:04
SIGAA | Secretaria de Tecnologia da Informação - STI - (61) 3107-0102 | Copyright © 2006-2024 - UFRN - app08.sigaa08