Aplicação de técnicas da inteligência artificial na seleção de ações para investimento na bolsa de valores de São Paulo

Abstract

The Stock Exchanges are institutions that mediate the buying and selling of stocks and securities, acting as a link between companies and investors. Investing in the Stock Exchanges has called the attention of new investors by providing greater financial returns compared to traditional forms of investment, such as the Fixed Income. However, this is a form of investment with a high degree of risk whereupon the investor should select a portfolio which combines maximized profits with minimized risk. Correctly identify trends in stock prices with the assistance of a technique is critical for this investor. Artificial Intelligence techniques can be applied in this identification. The objective of this work was to apply Artificial Intelligence techniques in selecting stocks to invest in the Stock Exchange of São Paulo, such as Artificial Neural Networks and the type Multilayer Perceptron Theory of Rough Sets. The experiments were performed in two stages divided into phases in order to select stocks for investment portfolios. Upon selection of portfolios we applied the Markowitz model and a Weighting Equation for risk management. The returns of managed portfolios were compared with the Bovespa index used as a benchmark. Obtained results appoint positively to the application of the two techniques of Artificial Intelligence in the selection of stock portfolios for investment on the Stock Exchange of São Paulo.