Run Monte Carlo Study for GOGARCH Estimation
run_gogarch_monte_carlo.RdPerforms a Monte Carlo simulation study to assess the accuracy of GOGARCH parameter estimation. Since GOGARCH estimates component-wise GARCH parameters, this evaluates estimation accuracy for each component.
Usage
run_gogarch_monte_carlo(
n_sim = 100,
n_obs = 500,
k = 3,
true_omega = NULL,
true_alpha = NULL,
true_beta = NULL,
omega = NULL,
alpha_garch = NULL,
beta_garch = NULL,
distribution = "norm",
shape = 8,
ica_method = "radical",
confidence_level = 0.95,
verbose = TRUE,
seed = 12345
)Arguments
- n_sim
Number of simulation replications
- n_obs
Number of observations per replication
- k
Number of series/components (default 3)
- true_omega
Vector of true component GARCH omega parameters (alias for omega)
- true_alpha
Vector of true component GARCH alpha parameters (alias for alpha_garch)
- true_beta
Vector of true component GARCH beta parameters (alias for beta_garch)
- omega
Vector of component GARCH omega parameters (alternative to true_omega)
- alpha_garch
Vector of component GARCH alpha parameters (alternative to true_alpha)
- beta_garch
Vector of component GARCH beta parameters (alternative to true_beta)
- distribution
Component distribution ("norm" or "std")
- shape
Degrees of freedom for "std" distribution
- ica_method
ICA algorithm ("radical" or "fastica")
- confidence_level
Confidence level for coverage (default 0.95)
- verbose
Print progress
- seed
Base seed for reproducibility
Value
List with:
- estimates
List of data frames (one per component) with alpha, beta columns
- persistence
List of vectors (one per component)
- convergence
Logical vector indicating optimization convergence
- ica_converged
Logical vector indicating ICA convergence
- bias
List of named vectors (one per component)
- rmse
List of named vectors (one per component)
- empirical_sd
List of named vectors (one per component)
- coverage
List of named vectors (one per component)
- mixing_recovery
Statistics on mixing matrix recovery
- summary
Summary data frame