Verification of Possibility of Forecasting Economical Time Series Using Neural Network and Digital Filter

This paper describes research that was performed to verify the possibility of learning an economical time series using a neural network. An actual economical time series wasn´t used for this research, but, instead, verification was performed using data obtained from an equation developed in the field of econophysics that simulates an economical time series. This data changes irregularly, and it is difficult to learn using a neural network, but by processing the data using a low pass digital filter, learning is possible with a low error rate, and short-term prediction is possible.