Improving the Criteria of the Investment on Stock Market Using Data Mining Techniques: The Case of S&P500 Index

Abstract

The stock market data, as S&P500 Index, is massive, complex, non-linear and noised. Thus, the investment criteria using this information have been a challenge. This study proposes the following short-term step by step strategy: to combine two information sources that the investors can analyse to make a decision. First, the index data constitutes the input for a Deep Learning Neural Network training, for representing and forecasting next day stock value. Second, this research identifies the most representative enterprises, included on Index, which represent the Index behavioural tendency, using Feature Selection Analysis. Finally, the outputs are complemented and corroborated; the process shows promising results to improve the investor's decision. Thus, the academics can revise a new experience in data analysis; for the practitioners, the research contributes to an approach for supporting investment decisions in the stock market.