Volatility Forecasting in Financial Risk Management with Statistical Models and ARCH-RBF Neural Networks

Abstract

As volatility plays very important role in financial risk management, we investigate the volatility dynamics of EUR/GBP currency. While a number of studies examines volatility using statistical models, we also use neural network approach. We suggest the ARCH-RBF model that combines information from ARCH with RBF neural network for volatility forecasting. We also use a large number of statistical models as well as different optimization techniques for RBF network such as genetic algorithms or clustering. Both insample and out-of-sample forecasts are evaluated using appropriate evaluation measures. In the final comparison none of the considered models performed significantly better than the rest with respect to the considered criteria. Finally, we propose upgrades of our suggested model for the future.