Mercado de capitais artificial : uma simulação baseada em agentes

• Ano de defesa: 2021
• Autor(a) principal: Castro, Conceição Jacqueline Xavier Barbosa de
• Orientador(a): Hadad Junior, Eli
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Presbiteriana Mackenzie
• Palavras-chave em Português: modelagem baseada em agentes; mercado de ações artificial; netlogo
• Área do conhecimento CNPq: CNPQ::CIENCIAS SOCIAIS APLICADAS::ADMINISTRACAO::ADMINISTRACAO DE EMPRESAS
• Link de acesso: https://dspace.mackenzie.br/handle/10899/28378

This dissertation aimed to compare and evaluate the predictive power of an artificial stock market model developed by agent-based modeling techniques (ABM). An empirical study was carried out to forecast the values of certain shares of B3, the Brazilian stock exchange, through the model of Collective Behavior in the Stock Market, created by Silva (2014). The values generated were compared with the reais collected between the dates: 06/05/2019 and the end date on 08/17/2020. The environment used was the Netlogo. The sample of data for comparison included 299 observations, with a simulation of 5400 models for each stock. The analysis comprised the comparison between the forecast and the actual closing price within established periods. The selection of models was performed using the Directional Accuracy test of Pesaran-Timmerman (2006). The results of the models were compared with a model based on statistics, using an ARIMA algorithm, by Box and Jenkins (1976). The models showed satisfactory initial results for the predictive power.