multiSA (version 0.1.0)

get_sdreport: Calculate standard errors

Description

A wrapper function to return standard errors. Various numerical techniques are employed to obtain a positive-definite covariance matrix.

Usage

get_sdreport(obj, getReportCovariance = FALSE, silent = FALSE, ...)

Value

Object returned by RTMB::sdreport(). A correlation matrix is generated and stored in: get_sdreport(obj)$env$corr.fixed

Arguments

obj

The list returned by RTMB::MakeADFun()

getReportCovariance

Logical, passed to RTMB::sdreport()

silent

Logical, whether to report progress to console. See details.

...

Other arguments to RTMB::sdreport() besides par.fixed, hessian.fixed, getReportCovariance

Details

In marginal cases where the determinant of the Hessian matrix is less than 0.1, several steps are utilized to obtain a positive-definite covariance matrix, in the following order:

1. Invert hessian returned by h <- obj@he(obj$env.last.par.best) (skipped in models with random effects)

2. Invert hessian returned by h <- stats::optimHess(obj$env.last.par.best, obj$fn, obj$gr)

3. Invert hessian returned by h <- numDeriv::jacobian(obj$gr, obj$env.last.par.best)

4. Calculate covariance matrix from chol2inv(chol(h))