SAS/SPSS Mixed Model cov type translation to Julia (MixedModels.jl)


I try to use MixedModels.jl.

Is any documentation how to translate SAS/SPSS random/repeated covariance type to julia MixedModels.jl notation?

for example:

RANDOM  Fact1/TYPE=FA0(2) SUB=Fact2 G;

where Fact1 and Fact2 - factors.

How to fit:

  • CSH (Heterogenous compound symmetry)
  • UN (Unstructured)
  • VC (Variance components)
  • AR1 (First-order autoregressive)

How to explain the relationship between these types and Julia MixedModels.jl notation to SAS users?


The model formula syntax that MixedModels.jl uses is largely (but not exactly) the same as the what the R package lme4 uses, so that might help you find some translation resources.