lme4-extractors: Extract Components from 'mmdml' Fits Imported from 'lme4'

In dmlalg: Double Machine Learning Algorithms

Description

Methods for the class mmdml for generics from lme4.

Usage

fixef(object, ...)
## S3 method for class 'mmdml'
fixef(object, ...)

ranef(object, ...)
## S3 method for class 'mmdml'
ranef(object, ...)

VarCorr(x, sigma = 1, ...)
## S3 method for class 'mmdml'
VarCorr(x, ...)

vcov(object, ...)
## S3 method for class 'mmdml'
vcov(object, ...)

Arguments

object, x	An object of class mmdml. This object usually results from a function call to mmdml.
sigma	See lmer from package lme4.
...	Further arguments passed to or from other methods.

Details

fixef.mmdml: Extracts the estimator of the linear coefficient β_0, which is a named and numeric vector.

ranef.mmdml: Extracts the random_eff entry from object.

VarCorr.mmdml: The variance and correlation components are computed with the sigma and the theta entries of x as in lmer. For each of the S repetitions, sigma and theta computed on the K sample splits are aggregated by taking the mean. Then, the S mean-aggregated estimates are aggregated by the median. The variance and correlation components are computed with these median-aggregated estimates.

vcov.mmdml: It returns the variance-covariance matrix of the estimator of the linear coefficient is extracted. It is computed based on the asymptotic Gaussian distribution of the estimator. First, for each of the S repetitions, the variance-covariance matrices computed on the K sample splits are aggregated by taking the mean. Second, the S mean-aggregated estimates are aggregated by adding a term correcting for the randomness in the sample splits and by taking the median of these corrected terms. This final corrected and median-aggregated matrix is returned.

Value

See lmer from package lme4.

See Also

mmdml

Examples

## See example(mmdml) for examples

dmlalg documentation built on Feb. 3, 2022, 5:08 p.m.