xrspatial.mcda.sensitivity.sensitivity#

xrspatial.mcda.sensitivity.sensitivity(criteria: Dataset, weights: dict[str, float], method: str = 'one_at_a_time', combine_method: str = 'wlc', delta: float = 0.05, n_samples: int = 1000, seed: int = 42, name: str = 'sensitivity') → DataArray | Dataset[source]#

Assess how weight changes affect the suitability surface.

Parameters:

  • criteria (xr.Dataset) – Standardized criterion layers.

  • weights (dict) – Base weights (must sum to ~1.0).

  • method (str) – "one_at_a_time" or "monte_carlo".

  • combine_method (str) – Combination function: "wlc" or "wpm".

  • delta (float) – Perturbation magnitude for one-at-a-time (default 0.05).

  • n_samples (int) – Number of random weight vectors for Monte Carlo (default 1000). Must be >= 1 when method="monte_carlo".

  • seed (int) – Random seed for Monte Carlo reproducibility.

  • name (str) – Name prefix for output arrays.

Returns:

For "one_at_a_time": Dataset with per-criterion DataArrays showing the absolute score difference under perturbation. For "monte_carlo": DataArray of per-pixel coefficient of variation across random weight samples.

Return type:

xr.Dataset or xr.DataArray