oraclesai.weights

class DistanceBandWeightsDefinition(threshold, p=2, alpha=-1.0, binary=True)

Defines a spatial relationship based on a given distance. It uses the distance between two nodes as the weight for the edge connecting them.

Parameters:

  • threshold – A distance.

  • p – float, default=2. Minkowski p-norm, where p=2 refers to the Euclidean distance, and p=1 refers to the Manhattan distance.

  • alpha – float, default=-1.0. Distance decay, if alpha is positive the weights won’t decline with distance. If binary is True, alpha is ignored.

  • binary – bool, default=True. If true, the weight \(w_{ij}=1\) if \(d_{ij}<=threshold\); otherwise, \(w_{ij}=0\). If false, \(w_{ij}=d_{ij}^{alpha}\).

class KNNWeightsDefinition(k: int = 4)

Defines a spatial relationship of K nearest neighbors.

Parameters:

k – Number of neighbors.

class KernelBasedWeightsDefinition(bandwidth=None, fixed=True, k=2, function='triangular')

Defines a spatial relationship based on a kernel function, a decay function where closer neighbors have larger values while further neighbors have smaller ones.

Parameters:

  • bandwidth – The distance used for the kernel function.

  • fixed – If True, the same bandwidth is used across all the records. Otherwise, the bandwidth is adaptive across observations.

  • k – Number of nearest neighbors used to calculate bandwidth. If fixed=True, the bandwidth is the maximum distance that guarantees that all observations have at least k neighbors; otherwise, the bandwidth is adaptive, so all observations have exactly k neighbors.

  • function – Name of the kernel function to be used. Available options are: triangular, uniform, quadratic, gaussian.

class QueenWeightsDefinition

Defines a spatial relationship where two polygons are considered neighbors if they share at least one vertex.

property index_allowed

Returns true if the spatial weights may be calculated by passing a spatial index of the data.

class RookWeightsDefinition

Defines a spatial relationship where two polygons are considered neighbors if they share at least one edge.

property index_allowed

Returns true if the spatial weights may be calculated by passing a spatial index of the data.

class SpatialWeights(w: W, definition: SpatialWeightsDefinition = None)
A SpatialWeights instance represents a spatial weights matrix calculated for some data

and a spatial neighboring relationship specified by an instance of SpatialWeightsDefinition.

as_pysal_weights() → W

Convert the current SpatialWeight object to its pysal representation

classmethod create(data, sw_def: SpatialWeightsDefinition)

Computes a SpatialWeights object for the given data.

Parameters:
Returns:

A SpatialWeights object.

property definition: SpatialWeightsDefinition

The class:oraclesai.weights.SpatialWeightsDefinition that describes the current spatial weights

classmethod from_pysal_weights(w: W)

Creates a SpatialWeight object from an existing pysal weights object

property neighbors: Dict

A dictionary containing the list of neighbors’ indexes for each observation

property weights: Dict

A dictionary containing the weights for each sample

class SpatialWeightsDefinition

A concrete implementation of this class defines the criteria used to identify spatial neighbors in a data set so that these neighbors can be used to calculate a spatial weights matrix.

classmethod from_pysal_weights(w)

Creates a SpatialWeightsDefinition that describes the given pysal weights

property index_allowed

Returns true if the spatial weights may be calculated by passing a spatial index of the data.