Commonly used in finance, this model is well suited for forecasting time series with volatility clustering properties
The Generalized Autoregressive Conditional Heteroscedastic model of order p,q, also known as GARCH (p,q), is a time series model that takes into account volatility, an important characteristic of financial data (e.g. volatility of asset returns). Forecasting volatility is useful in financial risk assessment.
The GARCH function implemented in XLSTAT-R calls the garch function of the tseries library (Adrian Trapletti, Kurt Hornik). It fits a GARCH model to time series by computing the maximum-likelihood estimates of the conditionally normal model.