estfun.gam: Estimating Robust/Clustered Standard Errors with 'mgcv'

In gKRLS: Generalized Kernel Regularized Least Squares

Estimating Robust/Clustered Standard Errors with mgcv

Description

This extracts the score of the log-likelihood for each observation for a model estimated using gam or bam. It is necessary to allow mgcv to work correctly with functions from sandwich to product cluster or robust standard errors.

Usage

## S3 method for class 'gam'
estfun(x, correct_df = TRUE, override_check = FALSE, ...)

Arguments

x	A model estimated using gam or bam.
correct_df	The default, TRUE, adjusts sandwich to use the effective degrees of freedom for k instead of the number of columns of the design. FALSE uses k.
override_check	The default, FALSE, allows this function to be used inside of sandwich. If only the matrix of scores is required, set to TRUE.
...	Not used for estfun.gam.

Value

A matrix used internally for sandwich.

gKRLS documentation built on Sept. 11, 2024, 8:24 p.m.