Detecção de comunidades em redes do mercado financeiro
Arquivos
Data
2021-08-11
Autores
Lopes, Renan Soares [UNIFESP]
Orientadores
Quiles, Marcos Gonçalves [UNIFESP]
Tipo
Trabalho de conclusão de curso de graduação
Título da Revista
ISSN da Revista
Título de Volume
Resumo
O mercado é um agrupamento adaptativo complexo, constituído por componentes inter-relacionados de agentes que interagem através do sistema de preços. Formalmente, pode ser descrito por uma rede dinâmica cuja topologia varia ao longo do tempo. O trabalho propõe modelar uma rede através da correlação entre os retornos históricos dos principais ativos que compõem o mercado brasileiro, representado pelo ETF BOVA11. Neste contexto, o intuito é analisar as características das comunidades --- grupos que compartilham funções e propriedades estruturais --- geradas pelo modelo. Além disso, também é feito um estudo sobre as crises econômicas que ocorreram no período analisado, observando as mudanças topológicas que alteram a estabilidade do sistema. Tal característica pode ser quantificada através da modularidade, uma medida estrutural da rede para determinar a qualidade da estrutura de comunidades. Com isso, espera-se contribuir com a discussão sobre a modelagem de problemas em redes financeiras, detecção de comunidades e crises econômicas expandindo a abordagem aplicando os conceitos em economias emergentes.
The market is a complex adaptive grouping, consisting of interrelated components of agents that interact through the price system. Formally, it can be described by a dynamic network whose topology varies over time. The work proposes to model a network through the correlation between the historical returns of the main assets that make up the Brazilian market, represented by the ETF BOVA11. In this context, the intention is to analyze the characteristics of the communities --- groups that share structural functions and properties --- generated by the model. In addition, a study is also made on the economic crises that occurred in the analyzed period, observing the topological changes that alter the stability of the system. Such a characteristic can be quantified through modularity, a structural measure of the network to determine the quality of the community structure. Thereby, it is expected to contribute to the discussion about the modeling of problems in financial networks, detection of communities and economic crises, expanding the approach applying the concepts in emerging economies.
The market is a complex adaptive grouping, consisting of interrelated components of agents that interact through the price system. Formally, it can be described by a dynamic network whose topology varies over time. The work proposes to model a network through the correlation between the historical returns of the main assets that make up the Brazilian market, represented by the ETF BOVA11. In this context, the intention is to analyze the characteristics of the communities --- groups that share structural functions and properties --- generated by the model. In addition, a study is also made on the economic crises that occurred in the analyzed period, observing the topological changes that alter the stability of the system. Such a characteristic can be quantified through modularity, a structural measure of the network to determine the quality of the community structure. Thereby, it is expected to contribute to the discussion about the modeling of problems in financial networks, detection of communities and economic crises, expanding the approach applying the concepts in emerging economies.