Nonstationary Generalised Autoregressive Conditional Heteroskedasticity Modelling for Fitting Higher Order Moments of Financial Series within Moving Time Windows

Figure 7

Here, we show how we take a rolling window for a long time series (from the start to the end). We highlight two fixed time windows, of certain length, T, with time median that corresponds to the middle of these windows. This time window then shifts in time, by , taken here to be six months.