Creates a sorted likelihood curve plot showing the performance of all calibration simulations. The plot displays simulations sorted by their log-likelihood values, with the best simulation highlighted.
Arguments
- results
A data frame containing calibration results with at minimum a 'likelihood' column, and optionally 'sim' and 'iter' columns for identifying individual simulations. This is typically the output from a model calibration run.
- output_dir
Character string specifying the directory where the plot should be saved. Directory will be created if it doesn't exist.
- verbose
Logical indicating whether to print messages. Default is TRUE.
Details
This function creates a publication-quality plot showing:
All simulations sorted by log-likelihood (x-axis: simulation index, y-axis: log-likelihood)
A smooth line connecting all points to show the curve
Individual points for each simulation
The best simulation highlighted with a red point
A horizontal dotted line at the maximum likelihood value
The plot helps visualize:
The overall calibration performance
The distribution of likelihood values
How many simulations achieved good fits
Whether there's a clear optimum or multiple good solutions
Examples
if (FALSE) { # \dontrun{
# Run calibration and collect results
results <- run_MOSAIC()
# Create likelihood curve plot
p_likelihood <- plot_model_likelihood(
results = results,
output_dir = "calibration_output"
)
# The plot is automatically saved, but you can also access it
print(p_likelihood)
} # }