predcond: Predicts means and variances conditionally on the factors

In factorstochvol: Bayesian Estimation of (Sparse) Latent Factor Stochastic Volatility Models

Predicts means and variances conditionally on the factors

Description

predcond simulates from the posterior predictive distribution of the data, conditionally on realized values of the factors. This has the advantage that the predictive density can be written as the product of the marginals but introduces sampling uncertainty that grows with the number of factors used.

Usage

predcond(x, ahead = 1, each = 1, ...)

Arguments

x	Object of class 'fsvdraws', usually resulting from a call to fsvsample.
ahead	Vector of timepoints, indicating how many steps to predict ahead.
each	Single integer (or coercible to such) indicating how often should be drawn from the posterior predictive distribution for each draw that has been stored during MCMC sampling.
...	Ignored.

Value

List of class fsvpredcond containing two elements:

means

Array containing the draws of the predictive means.

vols

Array containing the draws of the predictive volatilities (square root of variances).

See Also

Other predictors: predcor(), predcov(), predh(), predloglikWB(), predloglik(), predprecWB()

Examples

set.seed(1)
sim <- fsvsim(n = 500, series = 4, factors = 1) # simulate
res <- fsvsample(sim$y, factors = 1) # estimate

# Predict 1 day ahead:
predobj <- predcond(res, each = 5)

# Draw from the predictive distribution:
preddraws <- matrix(rnorm(length(predobj$means[,,1]),
                    mean = predobj$means[,,1],
                    sd = predobj$vols[,,1]), nrow = 4)

# Visualize the predictive distribution
pairs(t(preddraws), col = rgb(0,0,0,.1), pch = 16)

factorstochvol documentation built on Nov. 24, 2023, 5:08 p.m.