An empirical study of volatility predictions: Stock market analysis using neural networks

Abstract

Volatility is one of the major factor that causes uncertainty in short term stock market movement. Empirical studies based on stock market data analysis were conducted to forecast the volatility for the implementation and evaluation of statistical models with neural network analysis. The model for prediction of Stock Exchange short term analysis uses neural networks for digital signal processing of filter bank computation. Our study shows that in the set of four stocks monitored, the model based on moving average analysis provides reasonably accurate volatility forecasts for a range of fifteen to twenty trading days. © Springer-Verlag Berlin Heidelberg 2005.