RaschModels.jl is a Julia package for fitting and evaluating Rasch Models. It implements the basic Rasch Model, Partial Credit Model, and Rating Scale Model, as well as their linear extensions.
Note: Currently only a subset of models is available. Please see Roadmap for details.
To install this package you can use Julias package management system.
] add RaschModels
Fitting a model using RaschModels.jl is easy. First, get some response data as a Matrix
.
In this example we just use some random data for 100 persons and 5 items.
data = rand(0:1, 100, 5)
Using data
as our response data we can fit a Rasch Model.
rasch = fit(RaschModel, data, CML())
This function call fits the model using conditional Maximum Likelihood estimation. To fit the Rasch Model using Bayesian estimation just change the algorithm and provide the required additional arguments.
rasch_bayes = fit(RaschModel, data, NUTS(), 1_000)
Additional plotting capabilities are provided by ItemResponsePlots.jl.
RaschModels.jl is still under active development. Therefore, not all functionality is available yet. This roadmap provides a quick overview of the current state of the package.