cr.sparse.cluster.spectral.unnormalized_k

cr.sparse.cluster.spectral.unnormalized_k(key, W, k)[source]

Unnormalized spectral clustering with known number of clusters

Parameters
  • key – a PRNG key used for the k-means algorithm

  • W (jax.numpy.ndarray) – Similarity/Weights matrix

  • k (int) – The number of clusters

Returns

A named tuple with the spectral clustering solution (Laplacian, singular values, cluster assignment)

Return type

(SpectralclusteringSolution)