Métodos automatizados para investimento no mercado de ações via inteligência computacional

• Ano de defesa: 2017
• Autor(a) principal: Alexandre Pimenta
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Minas Gerais; UFMG
• Palavras-chave em Português: Computação evolucionária; Programação genética; Análise técnica; Otimização multiobjetivo; Mercado de ações; Programação genética (Computação); Bolsa de valores; Engenharia elétrica
• Link de acesso: http://hdl.handle.net/1843/BUOS-ATELWJ

Automated investment in sotck market is an area of strong interest of the academy and professional investors. In addition to conventional market methods, various sophisticated techniques have been employed to deal with the problem, such as ARCH/GARCH predictors, artificialneural networks, and fuzzy logic. This work presents two computational systems developed for this purpose. The first system is based on the optimal selection of rules of purchase and sale of shares. The second one combines conventional methods, genetic programming andmultiobjective optimization. Both systems aim to reach profit in the financial market, automatically identifying the best moments for purchasing and selling shares in a given time window. Unlike other types of investment systems, in which the objective is to try to predict the exactprice of the stock in the future, the systems developed in this work seek to identify suitable regions for purchase or sale of shares. The systems were tested in six historical time series of assets representative of the Brazilian stock market: BOVESPA. These assets are distributed infour distinct segments of the Brazilian economy. The systems presented higher gains than the valuation of the assets tested in the period.