Clustering algorithms using stochastic analysis and ensemble techniques.
from_noise(vecs, noise_map, ensemble_size=100, show_progress=False)
Given an array of vectors and a function which takes each vector to a randomly-generated ensemble, creates the ensemble of randomly-generated instances for each given vector. The output is a three-dimensional array
whose first dimension indexes the ensemble,
and whose remaining two dimensions have the same
shape as vecs
.
Arguments | Type | Description | |
---|---|---|---|
vecs |
np.ndarray |
A two-dimensional array, the first dimension of which indexes the vectors, and the remaining dimension is the dimension of the vectors. | |
noise_map |
function |
Any function which takes two positional arguments (a vector and an ensemble size) and generates a random ensemble of vectors (with ensemble index first). | |
ensemble_size |
Keyword | int |
The number of ensembles to run. |
show_progress |
Keyword | bool |
Whether or not to display a tqdm progress bar. |