htmd.clustering.kcenters module

class htmd.clustering.kcenters.KCenter(n_clusters)

Bases: BaseEstimator, ClusterMixin, TransformerMixin

Class to perform KCenter clustering of a given data set

KCenter randomly picks one point from the data, which is now the center of the first cluster. All points are put into the first cluster. In general the furthest point from its center is chosen to be the new center. All points, which are closer to the new center than the old one are assigned to the new cluster. This goes on, until K clusters have been created.

Parameters:

n_clusters (int) – desired number of clusters

Examples

>>> cluster = KCenter(n_cluster=200)
>>> cluster.fit(data)

cluster_centers

list with the points, which are the centers of the clusters

Type:

list

centerFrames

list of indices of center points in data array

Type:

list

labels_

list with number of cluster of each frame

Type:

list

clusterSize_

list with number of frames in each cluster

Type:

list

distance

list with the distance of each frame from the nearest center

Type:

list

fit(data)

Compute the centroids of data.

Parameters:

data (np.ndarray) – A 2D array of data. Columns are features and rows are data examples.