parglms-package {parglms} | R Documentation |
This package provides support for parallelized estimation of GLMs/GEEs, catering for dispersed data.
The DESCRIPTION file:
Package: | parglms |
Title: | support for parallelized estimation of GLMs/GEEs |
Version: | 1.25.1 |
Author: | VJ Carey <stvjc@channing.harvard.edu> |
Description: | This package provides support for parallelized estimation of GLMs/GEEs, catering for dispersed data. |
Suggests: | RUnit, sandwich, MASS, knitr, GenomeInfoDb, GenomicRanges, gwascat, BiocStyle, rmarkdown |
VignetteBuilder: | knitr |
Depends: | methods |
Imports: | BiocGenerics, BatchJobs, foreach, doParallel |
Maintainer: | VJ Carey <stvjc@channing.harvard.edu> |
License: | Artistic-2.0 |
LazyLoad: | yes |
BiocViews: | statistics, genetics |
ByteCompile: | TRUE |
git_url: | https://git.bioconductor.org/packages/parglms |
git_branch: | master |
git_last_commit: | 6771782 |
git_last_commit_date: | 2021-07-28 |
Date/Publication: | 2021-07-29 |
Index of help topics:
parGLM-methods fit GLM-like models with parallelized contributions to sufficient statistics parglms-package support for parallelized estimation of GLMs/GEEs
In version 0.0.0 we established an approach to fitting GLM from
data that have been persistently dispersed and managed by
a Registry
.
VJ Carey <stvjc@channing.harvard.edu>
Maintainer: VJ Carey <stvjc@channing.harvard.edu>
This package shares an objective with the bigglm
methods of biglm
. In bigglm
, a small-RAM-footprint algorithm
is employed, with sequential chunking to update statistics in each iteration.
In parGLM
the footprint is likewise controllable, but statistics
in each iteration are evaluated in parallel over chunks.
showMethods("parGLM")