MixedModels.jl

MixedModels.jl copied to clipboard

Extend rankUpdate! method to allow for upper tri

2

- Although this commit allows for `rankUpdate!` of the upper triangle it turns out that this is much slower than updating the lower triangle. ```julia julia> using BenchmarkTools, LinearAlgebra, MixedModels...

dmbates

raneftables returning NamedTuple of DictTables

2

- This is the first stage exploration of returning tabular representations for some of the `MixedModel` properties through `TypedTables.jl` - Change the result of `raneftables` to be a `NamedTuple` of...

dmbates

Change `hide_progress` in `parametricbootstrap` to `progress` and use `ProgressMeter`

1

Currently there is an inconsistency in both the argument name and its default between `fit!`, which has a `progress` argument defaulting to `true`, and `parametricbootstrap` which has a `hide_progress` argument...

dmbates

xref: https://github.com/JuliaStats/StatsModels.jl/pull/261

palday

Add TypedTables as a dependency to make it easier to return tables from simulations?

1

At present the allpars property of a parametric bootstrap sample is usually converted immediately to a DataFrame for manipulation. We probably don't want to have MixedModels.jl depend on DataFrames.jl but...

dmbates

Predict: only one level found

1

I wanted to `predict` a few values, but it said ```jl julia> MixedModels.predict(fitval, d) ERROR: ArgumentError: only one level found: 2016 (need at least two to compute contrasts). ``` My...

jariji

In the `Construction` section of the package documentation, it is said ```In contrast, the `cask` grouping factor is *nested* within the `batch` grouping factor in the *Pastes* data.``` I know...

likanzhan

If we want to have fitlog contain the initial and final parameter and objective values and have the `.initial`, `.final` fields just point to those values we should expand the...

palday

This happens e.g. when users have number of successes instead of proportion of successes.

palday

Recover from PIRLS failure by returning finitial

1

One potential downside to this approach is that it does make it possible to get stuck and/or have the optimizer believe that the initial values are the optimum. The former...

palday