Detect Parameter Sampling Flags from Config

Attempts to infer which parameters were originally sampled vs fixed by analyzing parameter variance across multiple configs. This is a fallback when original sampling flags are not available.

Usage

detect_sampled_parameters_from_configs(configs, variance_threshold = 1e-10)

Arguments

configs

List of config objects from parameter sampling

variance_threshold

Parameters with variance below this threshold are considered "fixed". Default 1e-10.

Value

Named list of logical values indicating which parameters appear to be sampled

Details

This function compares parameter values across multiple config objects to identify which parameters show variation (likely sampled) vs those that remain constant (likely fixed). This is a heuristic approach when original sampling metadata is lost.

Examples

if (FALSE) { # \dontrun{
# Generate multiple configs with some fixed parameters
configs <- lapply(1:10, function(i) {
  sample_parameters(PATHS, seed = i, sample_alpha_1 = FALSE, sample_phi_1 = TRUE)
})
sampling_flags <- detect_sampled_parameters_from_configs(configs)
} # }