pyriemann.clustering.Kmeans

class pyriemann.clustering.Kmeans(n_clusters=2, max_iter=100, metric='riemann', random_state=None, init='random', n_init=10, n_jobs=1, tol=0.0001)[source]

Clustering by k-means with SPD/HPD matrices as inputs.

The k-means is a clustering method used to find clusters that minimize the sum of squared distances between centroids and SPD/HPD matrices [1].

Then, for each new matrix, the class is affected according to the nearest centroid.

Parameters:
n_clustersint, default=2

Number of clusters.

max_iterint, default=100

Maximum number of iteration to reach convergence.

metricstring | dict, default=”riemann”

Metric used for mean estimation (for the list of supported metrics, see pyriemann.geometry.mean.gmean()) and for distance estimation (see pyriemann.geometry.distance.distance()). The metric can be a dict with two keys, “mean” and “distance” in order to pass different metrics.

random_stateNone | integer | np.RandomState, default=None

The generator used to initialize the centroids. If an integer is given, it fixes the seed. Defaults to the global numpy random number generator.

init“random” | ndarray, shape (n_clusters, n_channels, n_channels), default=”random”

Method for initialization of centroids. If “random”, it chooses k matrices at random for the initial centroids. If an ndarray is passed, it should be of shape (n_clusters, n_channels, n_channels) and gives the initial centroids.

n_initint, default=10

Number of time the k-means algorithm will be run with different centroid seeds. The final results will be the best output of n_init consecutive runs in terms of inertia.

n_jobsint, default=1

Number of jobs to use for the computation. This works by computing each of the n_init runs in parallel. If -1 all CPUs are used. If 1 is given, no parallel computing code is used at all, which is useful for debugging. For n_jobs below -1, (n_cpus + 1 + n_jobs) are used. Thus for n_jobs = -2, all CPUs but one are used.

tolfloat, default=1e-4

Stopping criterion to stop convergence, representing the minimum amount of change in labels between two iterations.

Attributes:
mdm_MDM instance

MDM instance containing the centroids.

labels_ndarray, shape (n_matrices,)

Labels, ie centroid indices, of each matrix of training set.

inertia_float

Sum of distances of matrices to their closest cluster centroids.

See also

Kmeans
MDM

Notes

Added in version 0.2.2.

References

__init__(n_clusters=2, max_iter=100, metric='riemann', random_state=None, init='random', n_init=10, n_jobs=1, tol=0.0001)[source]

Init.

centroids()[source]

Helper for fast access to the centroids.

Returns:
centroidsndarray, shape (n_clusters, n_channels, n_channels)

Centroids of each cluster.

fit(X, y=None)[source]

Fit the centroids of clusters.

Parameters:
Xndarray, shape (n_matrices, n_channels, n_channels)

Set of SPD/HPD matrices.

yNone

Not used, here for compatibility with sklearn API.

Returns:
selfKmeans instance

The Kmeans instance.

fit_predict(X, y=None)

Fit and predict in a single function.

Parameters:
Xndarray, shape (n_matrices, n_channels, n_channels)

Set of SPD matrices.

yNone

Not used, here for compatibility with sklearn API.

Returns:
predndarray of int, shape (n_matrices,)

Prediction for each matrix according to the closest cluster.

fit_transform(X, y=None, sample_weight=None)

Fit and transform in a single function.

Parameters:
Xndarray, shape (n_matrices, n_channels, n_channels)

Set of SPD matrices.

yNone | ndarray, shape (n_matrices,), default=None

Labels for each matrix.

sample_weightNone | ndarray, shape (n_matrices,), default=None

Weights for each matrix. If None, it uses equal weights.

Returns:
distndarray, shape (n_matrices, n_centroids)

Distance to each centroid.

get_metadata_routing()

Get metadata routing of this object.

Please check User Guide on how the routing mechanism works.

Returns:
routingMetadataRequest

A MetadataRequest encapsulating routing information.

get_params(deep=True)

Get parameters for this estimator.

Parameters:
deepbool, default=True

If True, will return the parameters for this estimator and contained subobjects that are estimators.

Returns:
paramsdict

Parameter names mapped to their values.

predict(X)[source]

Get the predictions.

Parameters:
Xndarray, shape (n_matrices, n_channels, n_channels)

Set of SPD/HPD matrices.

Returns:
predndarray of int, shape (n_matrices,)

Prediction for each matrix according to the closest centroid.

score(X, y, sample_weight=None)

Return the mean accuracy on the given test data and labels.

Parameters:
Xndarray, shape (n_matrices, n_channels, n_channels)

Test set of SPD matrices.

yndarray, shape (n_matrices,)

True labels for each matrix.

sample_weightNone | ndarray, shape (n_matrices,), default=None

Weights for each matrix.

Returns:
scorefloat

Mean accuracy of clf.predict(X) wrt. y.

set_output(*, transform=None)

Set output container.

See Introducing the set_output API for an example on how to use the API.

Parameters:
transform{“default”, “pandas”, “polars”}, default=None

Configure output of transform and fit_transform.

  • “default”: Default output format of a transformer

  • “pandas”: DataFrame output

  • “polars”: Polars output

  • None: Transform configuration is unchanged

Added in version 1.4: “polars” option was added.

Returns:
selfestimator instance

Estimator instance.

set_params(**params)

Set the parameters of this estimator.

The method works on simple estimators as well as on nested objects (such as Pipeline). The latter have parameters of the form <component>__<parameter> so that it’s possible to update each component of a nested object.

Parameters:
**paramsdict

Estimator parameters.

Returns:
selfestimator instance

Estimator instance.

set_score_request(*, sample_weight: bool | None | str = '$UNCHANGED$') → Kmeans

Configure whether metadata should be requested to be passed to the score method.

Note that this method is only relevant when this estimator is used as a sub-estimator within a meta-estimator and metadata routing is enabled with enable_metadata_routing=True (see sklearn.set_config()). Please check the User Guide on how the routing mechanism works.

The options for each parameter are:

  • True: metadata is requested, and passed to score if provided. The request is ignored if metadata is not provided.

  • False: metadata is not requested and the meta-estimator will not pass it to score.

  • None: metadata is not requested, and the meta-estimator will raise an error if the user provides it.

  • str: metadata should be passed to the meta-estimator with this given alias instead of the original name.

The default (sklearn.utils.metadata_routing.UNCHANGED) retains the existing request. This allows you to change the request for some parameters and not others.

Added in version 1.3.

Parameters:
sample_weightstr, True, False, or None, default=sklearn.utils.metadata_routing.UNCHANGED

Metadata routing for sample_weight parameter in score.

Returns:
selfobject

The updated object.

transform(X)[source]

Get the distance to each centroid.

Parameters:
Xndarray, shape (n_matrices, n_channels, n_channels)

Set of SPD/HPD matrices.

Returns:
distndarray, shape (n_matrices, n_clusters)

Distance to each centroid according to the metric.

Examples using pyriemann.clustering.Kmeans

Segmentation of hyperspectral image with Riemannian geometry

Segmentation of hyperspectral image with Riemannian geometry

Segmentation of SAR image with Riemannian geometry

Segmentation of SAR image with Riemannian geometry

Clustering algorithm comparison

Clustering algorithm comparison