Caracterização do mercado acionário brasileiro por meio dos cálculos de entropia, mutual information e a complexidade de Lempel-Ziv

We can describe financial markets as a Complex System, and the application of concepts from areas such as Econophysics and Information Theory provide a basis for understanding their characteristics and patterns. Thus, through the Entropy, Mutual Information (MI) and LempelZiv Complexity (CLZ) calcul...

ver descrição completa

Detalhes bibliográficos
Autores: Crepaldi, Antonio Fernando, Pacheco, Vinícius Henrique `Pinto
Formato: artículo
Estado:Versión publicada
Fecha de publicación:2021
País:Brasil
Recursos:Pontifícia Universidade Católica de São Paulo (PUC-SP)
Repositorio:Redeca
Idioma:portugués
OAI Identifier:oai:ojs.pkp.sfu.ca:article/56637
Acesso em linha:https://revistas.pucsp.br/index.php/redeca/article/view/56637
Access Level:acceso abierto
Palavra-chave:Sistemas Complexos
Econofísica
Mercados Financeiros
Entropia
Mutual Information
Complexidade de Lempel-Ziv
Complex Systems
Ecophysics
Financial Markets
Entropy
Lempel-Ziv Complexity
Descrição
Resumo:We can describe financial markets as a Complex System, and the application of concepts from areas such as Econophysics and Information Theory provide a basis for understanding their characteristics and patterns. Thus, through the Entropy, Mutual Information (MI) and LempelZiv Complexity (CLZ) calculations we sought to empirically characterize the series of daily stock returns on the São Paulo Stock Exchange between 2014 and 2018. From the results, it was seen that the most entropic days in the series were also those with variations in Ibovespa log returns and volume considerably higher than its daily average, reinforcing the fact that these are configured as days of less redundancy and greater informational content. For MI, the clusters were more delimited than those found via Pearson's Correlation. In CLZ, the low variability of substrings demonstrate a series with cyclic patterns and low complexity.