freediscovery.cluster.ClusterLabels¶

class freediscovery.cluster.ClusterLabels(vect, model, lsi=None, method='centroid-frequency', n_top_words=6)[source]¶

Calculate the cluster labels.

Parameters:

vect (VectorizerMixin object) – a scikit-learn’s text vectorizer
model (ClusterMixin object) – the cluster object
lsi_components (TruncatedSVD object or None) – LSA object if it was used for clustering
method (str, optional, default='centroid-frequency') – the method used to compute the centroid labels Only ‘centroid-frequency’ is supported at the moment.
n_top_words (int, default=10) – keep only most relevant n_top_words words

predict(centroids=None)[source]¶

Compute the cluster labels

Parameters:	centroids (list, default=None) – if not None, ignore clustering given by the clustering model and compute labels for the given cluster centroids
Returns:	cluster_labels
Return type:	array [n_samples]