Prediction of the RSD exchange rate by using wavelets and neural networks

Predicting the exchange rate of a certain currency on the global financial market of today plays an important role in economic calculations of every country and in its future investments. Hence, it is vital to control the prediction that is precise enough for this type of time series. This paper describes the system containing wavelets and one neural network that serves as a predictor for macroeconomic time series. The first part of the system utilizes the wavelet decomposition package and noise removal, while the second part performs the prediction of reconstructed signal aided by multi-layered feed forward neural network. The analysis of obtained results indicates that the suggested model sufficiently satisfies characteristics of the financial predictor.