bootinvFIM function

Approximation of the inverse of the Fisher Information Matrix via parametric bootstrap

When the FIM is not available, this function provides an approximation of the FIM based on an estimate of the covariance matrix of the model's parameters obtained via parametric bootstrap.

bootinvFIM(m, B = 1000, seed = 0)

Arguments

• m: a fitted model that will be used as the basis of the parametric bootstrap (providing the initial maximum likelihood estimate of the parameters and the modelling framework)
• B: the size of the bootstrap sample
• seed: a seed for the random generator

Returns

the empirical covariance matrix of the parameter estimates obtained on the bootstrap sample

Author(s)

Charlotte Baey <charlotte.baey@univ-lille.fr >