Abstract
This study applies artificial neural network to forecast Taiwan stock index option prices from 2005 to 2006. This study compares the option pricing model using three approaches to volatility: GARCH, EGARCH and GJR-GARCH, and compares estimators with three sets of criteria: RMSE, MAE and MAPE. The analytic results indicate that the asymmetric GARCH model is useful in predicting neural networks for stock index option prices.
Keywords and phrases: