An Algorithmic Trading Agent Based on a Neural Network Ensemble: A Case of Study in North American and Brazilian Stock Markets

In recent years, double auctions have become a topic of much interest in both economics and computer science literature, and automated trading has become very popular in stock markets. In this context, although many time series prediction studies are focused in the prediction of exact values in the future, evidences show that this kind of problem perform better when we transform it into a classification problem. In this work we propose a trading agent based on a neural network ensemble that predicts if one stock is going to raise or fall instead of predicting its future values. To show the efficiency of our method in different situations, we evaluate our model in two different datasets: the North American and the Brazilian stock markets. Real operations were simulated in these markets and we were able to profit in all tested time series.