Modelling of the History and Predictions of Financial Market Time Series Using Evolino

Abstract

Artificial neural networks and their systems are already capable of learning, to summarize, filter, and classify information. The increasing amount of authors are trying to teach them to approximate and predict chaotic, fractal processes. One of the greatest challenges of today's financial researches is forecasting of the commodities, stocks and currency markets. Variations in prices lead to economic indicators as result of investment process of investors and short time market players. Present article investigates recurrent neural network systems as mathematical tool for objective forecasts of fractal behaviour of financial markets by Evolino recurrent neural network learning algorithm.