SVR-wavelet adaptive model for forecasting financial time series

There is a necessity to anticipate and identify changes in events points to a new direction in the stock exchange markets in line with the analysis of the oscillations of prices of financial assets. This need leads to argue about new alternatives in the prediction of financial time series using machine learning methods. This paper aims to describe the development of the SVR-wavelet model, an adaptive and hybrid prediction model, which integrates wavelet models and Support Vector Regression (SVR), for prediction of financial time series, particularly applied to Foreign Exchange Market (FOREX), obtained from a public knowledge base. The method consists of using the Discrete Wavelets Transform (DWT) to decompose data from FOREX time series, that are used as SVR input variables to predict new data. The adjusted series are compared with traditional models such as ARIMA and ARFIMA Model. In Addition, statistical tests like normality and unit root are performed to prove that the series in question have non-linear distribution and also to verify the level of correlation between the periods of the series.