runLogLikelihoodEstimation
[Monolix] Log-likelihood estimation
Run the log-likelihood estimation algorithm. Note that the population
parameters must be already estimated (i.e. by calling runPopulationParameterEstimation). It is recommended
to call runConditionalModeEstimation first if using the linearization method.
The following methods are available:
| Method | Parameter | Log-Likelihood estimation by linearization |
linearization = TRUE | Log-Likelihood estimation by Importance Sampling (default) | linearization = FALSE |
The log-likelihood outputs(-2LL (OFV), AIC, BIC, BICc) are available using the getEstimatedLogLikelihood function.
Usage
runLogLikelihoodEstimation(linearization = FALSE)Arguments
- linearization
(logical) [optional]
TRUEto use linearization orFALSEto use stochastic approximation (the default)
See also
getEstimatedLogLikelihood to get the estimated log-likelihood runPopulationParameterEstimation to estimate population parameters runConditionalModeEstimation to estimate EBEs runConditionalDistributionSampling to estimate the conditional distributionrunStandardErrorEstimation to estimate standard errors of the population parametersrunScenario to run multiple estimation tasks
Examples
initializeLixoftConnectors("monolix")
loadProject(file.path(getDemoPath(), "1.creating_and_using_models", "1.1.libraries_of_models", "theophylline_project.mlxtran"))
runPopulationParameterEstimation()
runLogLikelihoodEstimation()
logLike <- getEstimatedLogLikelihood()