Prediction and Decision Making System Through Neural Networks for Investment Assets: Gold, Euro Dollar and Dow Jones

Abstract

The problem is to test the weak hypothesis of efficient markets through three neural networks that can predict the trends of investment assets such as: The Dow Jones, gold and Euro dollar, according to theories of technical analysis to automate positions of both long and short investment in the Spot market. With regard to forecasting time series, multiple approaches have been tested, through statistical models such as [1–3], where forecasts are made from different information sources with characteristics differentiated (sasonality, tendency, periodicity), however, other actors have begun to gain strength by getting the first places in international competitions, this is the case of Neural Networks, in works published as [4–6] the results have shown that this type of model offers a real opportunity to work with time series of different characteristics.