Sistema inteligente de apoio à tomada de decisão em mercado de ações: abordagem fundamentada em comportamento multitemporal

• Ano de defesa: 2017
• Autor(a) principal: Morais, Flávio Lopes de
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal de Lavras; Programa de Pós-Graduação em Engenharia de Sistemas e Automação; UFLA; brasil; Departamento de Engenharia
• Palavras-chave em Português: Bolsa de valores; Mercado de ações - Previsão; Inteligência computacional; Hipótese do Mercado Eficiente; Seguimento de tendência (Mercado); Bear markets; Stock price forecasting; Computational intelligence; Efficient Market Hypothesis; Trend following; Mercadologia; Sistemas de Informação
• Link de acesso: http://repositorio.ufla.br/jspui/handle/1/12829

Stocks correspond to a share of equity in a corporation, and they are traded mostly on stock exchanges by the process denominated secondary distribution. With the possibility of financial gains due to the constant changes in stock prices, speculative activities are a common practice in this market, leading to a highly competitive environment. The Efficient Market Hypothesis (EMH) assumes that it is not possible to obtain consistently excessive returns, since any type of information would be quickly reflected in share prices. In this work, a system model for speculative activities in the stock market was proposed, in order to verify the weak form of EMH. This model was proposed based on bibliographic studies and in a systematic literature review, conducted as part of this work. The proposal is based on the use of a trend-following strategy, based on multiple graphical times, supported by behavior prediction by the computational intelligence algorithms. The proposed model is based on behavioral characteristics of the stock market, originating from the Technical Analysis school and Dow Theory, and uses a set of technical indicators calculated from the own historical stock activity. In order to verify the feasibility of the proposal, an evaluation version of the system was developed. This version was then evaluated over a historical data from 30 shares of BM&FBovespa. On this database, the system was evaluated with an out-of-sample data that covered the whole period from 01/01/2010 to 08/31/2016. To a more realistic simulation, we considered transaction costs and bid-ask spread, as well as restrictions on the shares lot size. In the simulations, the proposed system achieved an average cumulative financial performance (ROI - Return On Investment), 20.32% higher than the market average. However, when considering the individual ROI on each share, the market performance was superior to the system in 53.33% of the cases, confirming the Efficient Market Hypothesis. Other performance measures indicate feasibility of the proposal, which have shown to be less risky and more likely to have positive financial results. The system was also compared to a Moving Average Crossovers strategy, achieving a ROI of 79.97% higher. To the same considered period in the simulations, while the IBOVESPA performance was -17.34%, the system achieved an average ROI of 97.47%. The results also showed the possibility of incre asing financial returns with the proposed system, since in the operation simulations the capital was used, on average, in 25.55% of the total available period. The proposal was also compared to the state of the art on the Shanghai Stock Exchange Composite Index, from 2001 to 2013. In this last comparison, the proposed system achieved an ROI 35.94% higher.