Uncertainty Quantification Model — create_uncertainty

Provides standardized data structure and validation for uncertainty quantification in ML-enhanced spatial sampling following constitutional principles. Supports epistemic, aleatoric, and total uncertainty types with proper validation and spatial consistency checks.

Usage

create_uncertainty_results(
  epistemic = NULL,
  aleatoric = NULL,
  total = NULL,
  uncertainty_rasters = NULL,
  confidence_intervals = NULL,
  mc_samples = NULL,
  n_samples = NULL,
  validation_metrics = NULL,
  method = "unknown",
  field_data = NULL
)

Arguments

epistemic: Epistemic (model) uncertainty estimates
aleatoric: Aleatoric (data) uncertainty estimates
total: Total uncertainty estimates (combined)
uncertainty_rasters: List of SpatRaster objects with uncertainty maps
confidence_intervals: List with confidence interval bounds
mc_samples: Monte Carlo samples for uncertainty estimation
n_samples: Number of Monte Carlo iterations used
validation_metrics: Uncertainty validation and calibration metrics
method: ML method used for uncertainty quantification
field_data: Reference field data for spatial consistency

Value

Standardized UncertaintyResults object

Details

UncertaintyResults ensures consistent representation of uncertainty estimates across all ML methods (BDL, RF, Ensemble) with constitutional compliance for spatial analysis excellence and performance requirements. Create standardized uncertainty quantification results

Examples

if (FALSE) { # \dontrun{
# Create uncertainty results
uncertainty_results <- create_uncertainty_results(
  epistemic = epistemic_uncertainty_raster,
  aleatoric = aleatoric_uncertainty_raster,
  total = total_uncertainty_raster,
  confidence_intervals = list(
    lower_bound = lower_ci_raster,
    upper_bound = upper_ci_raster,
    confidence_level = 0.95
  ),
  mc_samples = mc_sample_array,
  n_samples = 100,
  method = "BDL"
)
} # }