Applying Neural Networks in Copper Trading: Weekly Transaction ANN Model With a Linear Filter

摘要

An artificial neural network (ANN) is trained to generate buy/sell signals on a weekly basis for the copper market. The ANN model generates a signal that indicates what transaction shold be made one week in advance, establishing if the average price of copper for the next week will go up or down according to the average price of the present week. Velocity and acceleration of the rice and stock of copper and the best lienar estimation are used as inputs for the ANN model, which outperformes the martingale strategy by 8.8