adj_object: Loglikelihood adjustment of fitted objects

In RuoqingYin/oolax: Object-orientated Loglikelihood Adjustment for Extreme Value Models

Description

Performs loglikelihood adjustment of fitted objects for clustered data, based on Chandler and Bata (2007). We use vertical adjustment which is described in Section 6 in Chandler and Bata (2007).

Usage

adj_object(x, cluster = NULL, use_sandwich = TRUE, use_vcov = TRUE,
  ...)

Arguments

x	A fitted model object.
cluster	see details in adjust_loglik.
use_sandwich	A logical scalar. If use_sandwich = TRUE, we use meat or meatCL to estimate the V in adjust_loglik. Otherwise, set V as NULL.
use_vcov	A logical scalar. If use_vcov = TRUE and the method for x exists, we use it to estimate the Hessian of the independence loglikelihood and pass it as H to adjust_loglik. Otherwise, set H as NULL.
...	Further arguments to be passed to the functions in the sandwich package meat, if cluster = NULL, or meatCL, otherwise.

Details

The fitted object must have S3 methods: logLikVec, coef, and nobs. It may have method vcov and estfun. If a vcov method is not available then the variance-covariance matrix of the model parameters is estimated inside adjust_loglik. If an estfun method is not available then the score matrix is estimated using jacobian. More in See Also.

Value

An object of class c("oolax", "chandwich") with the same structure as an object returned from adjust_loglik.

See Also

adjust_loglik Loglikelihood adjustment using the sandwich estimator.

logLikVec.gev Loglikelihood vector for GEV Distribution fitted by fgev

logLikVec.pot Loglikelihood vector for GPD Distribution fitted by fpot

logLikVec.ismev_gev Loglikelihood vector for GEV Distribution fitted by gev.fit

logLikVec.ismev_gpd Loglikelihood vector for GPD Distribution fitted by gpd.fit

RuoqingYin/oolax documentation built on May 28, 2019, 12:20 p.m.