Four main RtIrt models are provided in this package,
GibbsMlIrtGibbsRtIrtGibbsRtIrtQuantile
These three models default to account for covariate variables (e.g., latent regression, and latent structure). If you need only a measurement model, you can use the null model.
GibbsRtIrtNull
You can download ExtendedRtIrtModeling directly from julia.
using Pkg
Pkg.add("ExtendedRtIrtModeling")or
] add ExtendedRtIrtModelingYou can also access the newest version and download it from GitHub.
using Pkg
Pkg.add(url="https://github.com/jiewenTsai/ExtendedRtIrtModeling.jl")or
] add "https://github.com/jiewenTsai/ExtendedRtIrtModeling.jl"Here is a simulation study example.
using ExtendedRtIrtModeling
## creat a toy data
Cond = setCond(nSubj=1000, nItem=15)
truePara = setTrueParaMlIrt(Cond)
Data = setDataMlIrt(Cond, truePara)
## build a model and sample it!
MCMC = GibbsMlIrt(Cond, Data=Data, truePara=truePara)
sample!(MCMC)
## check the parameter recovery
getRmse(MCMC.truePara.b, MCMC.Post.mean.b)
getBias(MCMC.truePara.b, MCMC.Post.mean.b)
If you have a data set to analyze, you can follow the following way,
using ExtendedRtIrtModeling
using CSV, DataFrames
## import your data set
yourData = CSV.read("yourData.csv", DataFrame)
Cond = setCond(qRa=0.85, qRt=0.85, nChain=3, nIter=3000)
Data = InputData(
Y=Matrix(yourData[:,1:15]),
T=exp.(Matrix(yourData[:,16:30])),
X=Matrix(yourData[:,31:33])
)
## build a model and sample it!
MCMC = GibbsRtIrtQuantile(Cond, Data=Data)
sample!(MCMC)
coef(MCMC)
precis(MCMC)
MCMC.Post.mean.Σp
MCMC.Post.mean.β
If you use ExtendedRtIrtModeling.jl in your work, please cite using the reference given in CITATION.cff.
If you want to make contributions of any kind, please first that a look into our contributing guide directly on GitHub or the contributing page on the website.