NLMEEstimation.jlEstimators for nonlinear mixed-effects models

Bring a conditional model; estimate random effects and population parameters.

Model protocol

PK comparison

Model-agnostic protocol

Use a per-subject likelihood or a batched program with automatic differentiation.

Empirical Bayes and FOCE(I)

Estimate subject effects at fixed population parameters, or fit the population by FOCE or FOCEI.

SAEM with MH or NUTS

Fit population parameters using stochastic approximation with per-subject MCMC transitions.

NLMEEstimation.jl defines estimation methods around a conditional likelihood supplied by a model. The package supplies the Gaussian random-effect prior and its covariance parameterization. Start with the model protocol, then compare the available estimators in the PK example.