Forecasting Stock Prices Volatility with Information (An ANN-GARCH Hybrid Approach)

Abstract

This study compares the forecast performance of volatilities between three models for forecasting stock returns: GARCH, hybrid ANN-GARCH with only GARCH output as the ANN input, and a hybrid ANN-GARCH with information. Through the extensive evaluation, the research found out that the hybrid ANN-GARCH model with information outperforms the other two models in terms of forecasting accuracy and predictive power. This study is set to find out the improvement performance of the hybrid ANN-GARCH with information vis a vis the Univariate GARCH Keywords: Stock price forecasting, GARCH, Artificial Neural Network DOI: 10.7176/RJFA/14-17-04 Publication date: September 30 th 2023