MixedModels.jl
MixedModels.jl copied to clipboard
JuliaStats
A Julia package for fitting (statistical) mixed-effects models
Nothing like a good visualization to make everything clear.
For those times when you only care about a single blocking variable. cc @kleinschmidt
Copying some discussion from the Julia Slack. @dmbates wrote: > We have recently begun working with a large data set to stress-test the MixedModels.jl package. Essentially all the time to...
We want this to work again: ```julia using MixedModels cbpp = MixedModels.dataset(:cbpp) gm = fit(MixedModel, @formula((incid/hsz) ~ 1 + period + (1|herd)), cbpp, Binomial(), wts=float(cbpp.hsz)) predict(gm, cbpp; type=:response) ≈ fitted(gm)...
For historical reasons, the bootstrap and `replicate` use `hide_progress=false` as the default, while the fitting functions use `progress=true`. We should change everything to be `progress` to **show** the progress.
We can do something using Arrow, but reducing precision to Float32 or even Float16. This would complement `saveoptsum`.
And probably also the parallel for `raneftables`, so that we can take full advantage of #545 downstream
- [x] dof_residual(model::RegressionModel) - [x] fitted(model::RegressionModel) - [x] modelmatrix(model::RegressionModel) - [x] predict(model::RegressionModel) - [x] residuals(model::RegressionModel) - [x] response(model::RegressionModel) - [ ] cooksdistance(model::RegressionModel) - [ ] crossmodelmatrix(model::RegressionModel) - [x] leverage(model::RegressionModel) -...
This is following up on a [post](https://discourse.julialang.org/t/mixed-multi-membership-models-in-julia/42005) on the Julia discourse board. In several mixed modeling packages in R, it is possible to fit multi-membership models where each row in...
This is discussed in lots of places elsewhere, but maybe we can consolidate our take and our priorities into a page like the one for rank deficiency.
Metadata
Owner
Metadata
A Julia package for fitting (statistical) mixed-effects models