Modeling volatility of time series using fuzzy GARCH models

Fuzzy modeling is an effective way of construction models for complex dynamic systems. Here we present a new application of fuzzy rule-based models to analysis of discrete time series. Fuzzy generalization of autoregressive conditional heteroscedasticity (ARCH/GARCH) models is proposed and technology of fuzzy GARCH modeling is basically overviewed. A comparison with usual GARCH models is made both for modeled and real time series.