shrinkTVPVAR (version 0.1.1)

simTVPVAR: Generate synthetic data from a TVP-VAR-SV model

Description

simTVPVAR generates synthetic data from a TVP-VAR-SV model. The data is always generated as to be stationary. This is done via a trial and error approach, where the VAR coefficients are drawn from the data generating process until the VAR process is stationary. As such, very large models might take a long time to generate.

Usage

simTVPVAR(
  N = 200,
  p = 2,
  m = 3,
  prob_0_beta = 0.8,
  prob_0_theta = 0.8,
  simsig2_theta_sr = 0.2,
  simsig2_beta_mean = 0.2,
  intercept = TRUE,
  display_progress = TRUE
)

Value

The value returned is a list object containing:

• data: data frame that holds the simulated data.

• true_vals: list object containing:

  • Phi: array containing the true VAR coefficients.

  • Sigma: array containing the true covariance matrices.

  • theta_sr: array containing the true standard deviations of the theta matrix.

  • beta_mean: array containing the true means of the beta matrix.

Arguments

N

integer > 2. Indicates the length of the time series to be generated. The default value is 200.

p

integer > 0. Indicates the number of lags in the VAR model. The default value is 2.

m

integer > 1. Indicates the number of equations in the VAR model. The default value is 3.

prob_0_beta

numeric. Indicates the probability of a zero element in the beta_mean matrix. Can be a single value or a vector of length p. The default value is 0.8.

prob_0_theta

numeric. Indicates the probability of a zero element in the theta matrix. Can be a single value or a vector of length p. The default value is 0.8.

simsig2_theta_sr

numeric. Indicates the standard deviation of the normal distribution from which the elements of the theta matrix are drawn. The default value is 0.2.

simsig2_beta_mean

numeric. Indicates the standard deviation of the normal distribution from which the elements of the beta_mean matrix are drawn. The default value is 0.2.

intercept

logical. Indicates whether an intercept should be included in the model. The default value is TRUE.

display_progress

logical. Indicates whether a progress bar should be displayed. The default value is TRUE.

Author

Peter Knaus

Examples

# \donttest{
# Generate a time series of length 300
res <- simTVPVAR(N = 300, m = 3, p = 3)

# Estimate a model
model <- shrinkTVPVAR(y = res$data, p = 3)
# }