ForecastHMMES: Expected shortfall (ES) of a univariate HMM at time n+k1,...

In GenHMM1d: Goodness-of-Fit for Univariate Hidden Markov Models

Description

This function computes the ES of a univariate HMM for multiple horizons, given observations up to time n

Usage

ForecastHMMES(U, family, theta, Q, eta, k = 1)

Arguments

U	value (1 x 1) at which the expected shortfall needs to be computed; between 0 and 1; (e.g 0.01, 0.05)
family	distribution name; run the function distributions() for help
theta	parameters; (r x p)
Q	probability transition matrix; (r x r)
eta	vector of the estimated probability of each regime at time n; (1 x r)
k	prediction times (may be a vector of integers).

Value

es	expected shortfall (1 x horizon)

Examples

family = "gaussian"

theta = matrix(c(-1.5, 1.7, 1, 1),2,2)
Q = matrix(c(0.8, 0.3, 0.2, 0.7), 2, 2)
eta = c(0.96091218, 0.03908782)

forecastedES = ForecastHMMES(U=c(0.01), family, theta=theta, Q=Q, eta=eta, k=c(1,2,3,4,5))
print('Forecasted expected shortfall : ')
print(forecastedES)

GenHMM1d documentation built on Jan. 21, 2021, 9:07 a.m.