spateo.alignment.methods.sampling¶

Classes¶

TRNET

Class for topology representing network sampling.

Functions¶

`sample`(→ numpy.ndarray)	A collection of various sampling methods.
`trn`(→ numpy.ndarray)	Sample method based on topology representing network.
`sample_by_velocity`(→ numpy.ndarray)	Sample method based on velocity.
`sample_by_kmeans`(→ Optional[numpy.ndarray])	Sample method based on kmeans.
`lhsclassic`(→ numpy.ndarray)	Latin Hypercube Sampling method implemented from PyDOE.

Module Contents¶

spateo.alignment.methods.sampling.sample(arr: list | numpy.ndarray, n: int, method: Literal['random', 'velocity', 'trn', 'kmeans'] = 'random', X: numpy.ndarray | None = None, V: numpy.ndarray | None = None, seed: int = 19491001, **kwargs) → numpy.ndarray[source]¶

A collection of various sampling methods.

Parameters:

arr: The array to be sub-sampled.
n: The number of samples.
method: The method to be used. “random”: randomly choosing n elements from arr; “velocity”: Higher the velocity, higher the chance to be sampled; “trn”: Topology Representing Network based sampling; “kmeans”: n points that are closest to the kmeans centroids on X are chosen. Defaults to “random”.
X: Coordinates associated to each element in arr. Defaults to None.
V: Velocity associated to each element in arr. Defaults to None.
seed: The randomization seed. Defaults to 19491001.

Raises:

NotImplementedError – method is invalid.

Returns:

The sampled data array.

class spateo.alignment.methods.sampling.TRNET(n_nodes: int, X: numpy.ndarray, seed: int = 19491001)[source]¶

Class for topology representing network sampling.

n_nodes[source]¶: The number of nodes in the graph.

n_dim¶: The dimensions of the array to be sub-sampled.

X[source]¶: Coordinates associated to each element in original array to be sub-sampled.

seed[source]¶: The randomization seed.

W[source]¶: The sample graph.

n_nodes[source]¶

n_dims[source]¶

X[source]¶

seed = 19491001[source]¶

W[source]¶

draw_sample(n_samples: int) → numpy.ndarray[source]¶

Initialize the positions of nodes.

Parameters:

n_samples: The number of nodes.

Returns:

The initial positions of nodes.

runOnce(p: numpy.ndarray, l: float, ep: float, c: float) → None[source]¶

Performs one iteration of the TRN sampling algorithm. Learn from distance to update the sampling index.

Parameters:

p: The target data points to calculate the distance.
l: The learning rate that controls learning speed.
ep: The epsilon that controls learning speed.
c: The cutoff parameter to accelerate the learning.

run(tmax: int = 200, li: float = 0.2, lf: float = 0.01, ei: float = 0.3, ef: float = 0.05, c: float = 0) → None[source]¶

Runs the TRN sampling algorithm for the specified number of iterations.

Parameters:

tmax: The maximum number of iterations.
li: The initial learning rate parameter.
lf: The final learning rate parameter.
ei: The initial epsilon parameter.
ef: The final epsilon parameter.
c: The cutoff parameter to accelerate the learning.

run_n_pause(k0: int, k: int, tmax: float = 200, li: float = 0.2, lf: float = 0.01, ei: float = 0.3, ef: float = 0.05, c: int = 0) → None[source]¶

Run the TRN algorithm sampling for a specified range of iterations.

Parameters:

k0: Starting iteration number.
k: Ending iteration number.
tmax: The maximum number of iterations.
li: The initial learning rate parameter.
lf: The final learning rate parameter.
ei: The initial epsilon parameter.
ef: The final epsilon parameter.
c: The cutoff parameter to accelerate the learning.

spateo.alignment.methods.sampling.trn(X: numpy.ndarray, n: int, return_index: bool = True, seed: int = 19491001, **kwargs) → numpy.ndarray[source]¶

Sample method based on topology representing network.

Parameters:

X: Coordinates associated to each element in original array to be sub-sampled.
n: The number of samples.
return_index: Whether to return the indices of the sub-sampled array or the sample graph. Defaults to True.
seed: The randomization seed. Defaults to 19491001.

Returns:

The sample graph or the indices of the sub-sampled array.

spateo.alignment.methods.sampling.sample_by_velocity(V: numpy.ndarray, n: int, seed: int = 19491001) → numpy.ndarray[source]¶

Sample method based on velocity.

Parameters:

V: Velocity associated with each element in the sample array.
n: The number of samples.
seed: The randomization seed. Defaults to 19491001.

Returns:

The sample data array.

spateo.alignment.methods.sampling.sample_by_kmeans(X: numpy.ndarray, n: int, return_index: bool = False) → numpy.ndarray | None[source]¶

Sample method based on kmeans.

Parameters:

X: Coordinates associated to each element in arr.
n: The number of samples.
return_index: Whether to return the sample indices. Defaults to False.

Returns:

The sample index array if return_index is True. Else return the array after sampling.

spateo.alignment.methods.sampling.lhsclassic(n_samples: int, n_dim: int, bounds: numpy.ndarray | List[List[float]] = None, seed: int = 19491001) → numpy.ndarray[source]¶

Latin Hypercube Sampling method implemented from PyDOE.

Parameters:

n_samples: The number of samples to be generated.
n_dim: The number of data dimensions.
bounds: n_dim-by-2 matrix where each row specifies the lower and upper bound for the corresponding dimension. If None, it is assumed to be (0, 1) for every dimension. Defaults to None.
seed: The randomization seed. Defaults to 19491001.

Returns:

The sampled data array.