Forecasting Stock Price Volatility Incorporating Covariate Information (An ANN-GARCH Hybrid Approach)

Abstract

Volatility forecasting in the financial markets is important in the areas of risk management and asset pricing.GARCH models are widely used in forecasting volatile time series data.The errors in prediction when using this approach are often quite high.Therefore,this study seeks to improve the performance of GARCH models by using artificial neural networks. The motivation of this study is to decide whether a hybrid model with additional information can improve the stocks volatility forecasts and by what percentage.The main objective of this study is to model stock prices volatilities using hybrid ANN- GARCH with additional information and compare the result to the hybrid ANN-GARCH and standalone GARCH.