predict_unconditional_quantiles: Predict unconditional extreme quantiles using peaks over...

In EQRN: Extreme Quantile Regression Neural Networks for Risk Forecasting

Predict unconditional extreme quantiles using peaks over threshold

Description

Predict unconditional extreme quantiles using peaks over threshold

Usage

predict_unconditional_quantiles(interm_lvl, quantiles = c(0.99), Y, ntest = 1)

Arguments

interm_lvl	Probability level at which the empirical quantile should be used as the intermediate threshold.
quantiles	Probability levels at which to predict the extreme quantiles.
Y	Vector of ("training") observations.
ntest	Number of "test" observations.

Value

Named list containing:

predictions	matrix of dimension ntest times length(quantiles) containing the estimated extreme quantile at levels quantile, repeated ntest times,
pars	matrix of dimension ntest times 2 containing the two GPD parameter MLEs, repeated ntest times.
threshold	The threshold for the peaks-over-threshold GPD model. It is the empirical quantile of Y at level interm_lvl, i.e. stats::quantile(Y, interm_lvl).

EQRN documentation built on April 4, 2025, 12:45 a.m.