Previsão de valor de um ativo financeiro utilizando redes neurais artificiais para processamento de dados de séries temporais

Abstract

The prediction of the Time Series(ST) plays a major role in the economy, especially in the branch of investment based on the purchase and sale of stocks of publicly traded companies, where it is very important to predict if the stock price will rise or fall, in order to maximize profits or minimize losses. A time series is a set of observations on a data type, ordered in time and recorded in regular periods. The basic assumption leading to time-series analysis is that there is a time-related system that exerted influence on data in the past and possibly will continue to exercise in the future, giving rise to non-random patterns, which may or may not be interpreted by graphs , Statistical and mathematical processes or artificial neural networks. The most common models for predicting the financial market are mathematical or statistical and depend on advanced technical knowledge and many model adaptations.