pymixef.backends.lmm module

Dense Gaussian linear mixed-model reference engine.

This engine evaluates the exact marginal Gaussian likelihood and restricted likelihood. It supports multiple independent random-effect blocks, correlated or diagonal within-group covariance, known observation weights, and an explicit structured residual covariance template. It deliberately forms dense covariance matrices and is therefore a correctness/reference engine rather than a sparse high-scale production engine.

pymixef.backends.lmm.DenseLMMBackend

alias of GaussianLMMBackend

class pymixef.backends.lmm.GaussianLMMBackend

Bases: object

Exact dense/reference Gaussian LMM engine.

name = 'lmm'

fit(data, *, method=None, reml=None, maxiter=1_000, tolerance=1e-8, compute_hessian=True, **options)

Parameters:

• data (Any)

• method (str | None)

• reml (bool | None)

• maxiter (int)

• tolerance (float)

• compute_hessian (bool)

• options (Any)

Return type:

dict[str, Any]

pymixef.backends.lmm.LMMBackend

alias of GaussianLMMBackend

pymixef.backends.lmm.fit_lmm(data, **options)

Fit a Gaussian LMM with the dense reference backend.

Parameters:

• data (Any)

• options (Any)

Return type:

dict[str, Any]