| Title: | Variance-Covariance Matrices and Standard Errors |
|---|---|
| Description: | Methods for faster extraction (about 5x faster in a few test cases) of variance-covariance matrices and standard errors from models. Methods in the 'stats' package tend to rely on the summary method, which may waste time computing other summary statistics which are summarily ignored. |
| Authors: | Michael Chirico |
| Maintainer: | Michael Chirico <[email protected]> |
| License: | GPL (>= 2) | file LICENSE |
| Version: | 0.0.1 |
| Built: | 2024-07-06 02:42:13 UTC |
| Source: | https://github.com/michaelchirico/vcov |
This package is designed to produce variance-covariance matrices and standard errors as directly/efficiently as possible from fit models. Default methods (e.g., in stats) tend to first compute the summary object for a model, from which the matrix is extracted. The catch is that the summary itself often involves several other extraneous computations. The summary methods are typically fast for most purposes, but falter in this regard when a user wishes to compute standard errors of perhaps thousands of models (as may happen, for example, when bootstrapping).
Runs a set of tests to check vcov is working correctly.
test.vcov(...)test.vcov(...)
... |
Currently no arguments. |
Runs a series of tests.
test.vcov()test.vcov()
Skip wasted object summary steps computed by base R when computing covariance matrices and standard errors of common model objects.
Vcov(object, ...) ## S3 method for class 'lm' Vcov(object, ...) ## S3 method for class 'glm' Vcov(object, dispersion = NULL, ...) se(object, ...)Vcov(object, ...) ## S3 method for class 'lm' Vcov(object, ...) ## S3 method for class 'glm' Vcov(object, dispersion = NULL, ...) se(object, ...)
object |
A fitted model object. |
... |
Additional arguments for method functions. For the |
dispersion |
The dispersion parameter for the family used. Either a single numerical value or |
# data taken from ?lm ctl = c(4.17,5.58,5.18,6.11,4.50,4.61,5.17,4.53,5.33,5.14) trt = c(4.81,4.17,4.41,3.59,5.87,3.83,6.03,4.89,4.32,4.69) group = gl(2, 10, 20, labels = c("Ctl","Trt")) weight = c(ctl, trt) reg_lm = lm(weight ~ group) Vcov(reg_lm) se(reg_lm) # data taken from ?glm counts = c(18,17,15,20,10,20,25,13,12) outcome = gl(3,1,9) treatment = gl(3,3) reg_glm = glm(counts ~ outcome + treatment, family = poisson) Vcov(reg_glm) se(reg_glm)# data taken from ?lm ctl = c(4.17,5.58,5.18,6.11,4.50,4.61,5.17,4.53,5.33,5.14) trt = c(4.81,4.17,4.41,3.59,5.87,3.83,6.03,4.89,4.32,4.69) group = gl(2, 10, 20, labels = c("Ctl","Trt")) weight = c(ctl, trt) reg_lm = lm(weight ~ group) Vcov(reg_lm) se(reg_lm) # data taken from ?glm counts = c(18,17,15,20,10,20,25,13,12) outcome = gl(3,1,9) treatment = gl(3,3) reg_glm = glm(counts ~ outcome + treatment, family = poisson) Vcov(reg_glm) se(reg_glm)