Plot Posterior Parameter Correlation Heatmap — plot_model_parameter

Generates a heatmap of pairwise Spearman correlations between parameters in the posterior (best subset), revealing trade-offs and redundancies. Parameters are clustered hierarchically and only correlations exceeding a threshold are labeled.

Usage

plot_model_parameter_correlation(
  results_file,
  priors_file = NULL,
  output_dir = ".",
  cor_threshold = 0.3,
  max_params = 25,
  subset_col = "is_best_subset",
  verbose = TRUE
)

Arguments

results_file: Path to samples.parquet file containing calibration results
priors_file: Path to priors.json (used for parameter descriptions)
output_dir: Directory to write the output figure
cor_threshold: Minimum absolute correlation to display as text label. Default 0.3.
max_params: Maximum number of parameters to display. The top max_params most-correlated parameters are selected. Default 25.
subset_col: Character name of the boolean subset-membership column. Defaults to "is_best_subset". Pass "is_best_subset_opt" to read the optimizer-refined subset.
verbose: Print progress messages

Value

Invisible NULL. Writes a PNG file to output_dir.