Estimation of the Normal-Inverse Gaussian Distribution

Performs maximum likelihood estimation on the Normal-Inverse Gaussian (NIG) distribution (univariate and multivariate). Wrappers fit.NIGuv and fit.NIGmv.

fit_nig(x, symmetric = FALSE)

# S3 method for default
fit_nig(x, symmetric = FALSE)

# S3 method for tbl
fit_nig(x, symmetric = FALSE)

# S3 method for xts
fit_nig(x, symmetric = FALSE)

# S3 method for matrix
fit_nig(x, symmetric = FALSE)

Arguments

x	A tabular (non-tidy) data structure.
symmetric	A logical flag. Should the estimated distribution be symmetric? Defaults to FALSE.

Value

A list of the the class cma_fit with 21 components.

See also

fit_ghd fit_hyp fit_vg fit_t fit_normal

Examples

x <- matrix(diff(log(EuStockMarkets)), ncol = 4)

# multivariate estimation
fit_nig(x)
#> # Margins Estimation
#> Converged:       TRUE
#> Dimension:       4
#> AIC:            -52708.2
#> Log-Likelihood:  26373.1
#> Model:           Asymmetric Normal Inverse Gaussian

# univariate estimation
fit_nig(x[ , 4, drop = FALSE])
#> # Margins Estimation
#> Converged:       TRUE
#> Dimension:       1
#> AIC:            -12786.8
#> Log-Likelihood:  6397.4
#> Model:           Asymmetric Normal Inverse Gaussian