Functional time series forecasting of extreme values

We consider forecasting functional time series of extreme values within a generalized value distribution (GEV). The GEV can be characterized using the three parameters (location, scale, and shape). As result, forecasts density accomplished by these latent parameters. Depending on underlying data structure, some either modeled as scalars or functions. provide two algorithms to model forecast To assess uncertainty, we apply sieve bootstrap method construct pointwise simultaneous prediction intervals forecasted values. Illustrated daily maximum temperature dataset, demonstrate advantages modeling Further, finite-sample performance our methods is quantified several Monte Carlo simulated under range scenarios.