Qualitative Analysis of Artificial Neural Networks Applied to Forecast Trends in the Brazilian Stock Market
Gustavo José Corrêa Gonçalves1; Eduardo Paciência Godoy1
1 UNESP - Sorocaba
doi:10.20906/CPS/COB-2015-0335
Resumo
The interest over the stock market has increased in the last decade, once it can offer considerable profits, from very short to long term. However, the stocks graphical analysis is not an exact technique, so subjective is present in it, depending on the person who performs the analysis. Thus, the use of Artificial Neural Networks is prominent when applied to predict movements of stocks traded on the stock exchange. This work proposes the use of Nonlinear Autoregressive Neural Networks with Exogenous Inputs to predict trends over daily close price of preferred shares of Itaú Bank. Furthermore, the influences of configuration parameters of the network over the predictions are analyzed.
Palavras-chave: Artificial Neural Networks; NARX; Stock Market