pygeostats.spatial_autocorrelation¶
Spatial autocorrelation methods.
global_getis_ord_g(values: ndarray, weights: ndarray, permutations: int = 0, random_state: Optional[int] = None) -> Dict[str, float]
Compute global Getis-Ord General G statistic.
This implementation returns the observed G and an optional permutation p-value around the permutation null distribution.
local_getis_ord_g(values: ndarray, weights: ndarray, include_self: bool = True) -> Dict[str, np.ndarray]
Compute local Getis-Ord G_i* z-scores.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
values
|
(ndarray, shape(n))
|
Observed attribute values. |
required |
weights
|
(ndarray, shape(n, n))
|
Spatial weights matrix. |
required |
include_self
|
bool
|
Whether to include self-weight in each local neighborhood. |
True
|
gearys_c(values: ndarray, weights: ndarray, permutations: int = 0, random_state: Optional[int] = None) -> Dict[str, float]
Compute global Geary's C with optional permutation p-value.
Compute local Geary components for each observation.
Compute local Moran's I for each observation.
morans_i(values: ndarray, weights: ndarray, permutations: int = 0, random_state: Optional[int] = None) -> Dict[str, float]
Compute global Moran's I with optional permutation p-value.
Row-standardize a square weights matrix.
spatial_weights_distance_band(coords: ndarray, threshold: float, binary: bool = True, row_standardize: bool = True) -> np.ndarray
Build a distance-band spatial weights matrix.
If binary is False, uses inverse distance weights inside the threshold.
spatial_weights_inverse_distance(coords: ndarray, power: float = 1.0, max_distance: Optional[float] = None, row_standardize: bool = True) -> np.ndarray
Build an inverse-distance spatial weights matrix.
Build a binary k-nearest-neighbor spatial weights matrix.