This R package provides an implementation of the GEE method for estimating multinomial probabilties for clustered data, for independent and exchangeable correlation structure.
Landsman V., Landsman D., Li C.S., Bang H. Overdispersion models for correlated multinomial data: Applications to blinding assessment. Statistics in Medicine. 2019; 38: 4963– 4976. https://doi.org/10.1002/sim.8344
There are a few methods available for installing the package. You need to have the devtools
package installed to use these methods:
> install.packages('devtools')
Use the install_github
method from devtools to install directly from this repository:
> devtools::install_github('https://github.com/Express50/omgee')
Download the latest released .tar.gz file from the Releases page into an empty folder. Then, navigate to the location of the file in your command line, and run the following from an R shell:
> devtools::install('.', dependencies=TRUE)
You can also build the code yourself by cloning the repository and using R CMD BUILD
. Make sure to install the rootSolve
dependecy before installing omgee
, by running install.packages('rootSolve')
> git clone https://github.com/Express50/omgee omgee
> R CMD build omgee
> R CMD INSTALL omgee_1.0.tar.gz