classix.merging.density_merging

classix.merging.density_merging(data, splist, radius, sort_vals, half_nrm2)

Implement CLASSIX’s merging with disjoint-set data structure, default choice for the merging.

Parameters

datanumpy.ndarray

The input that is array-like of shape (n_samples,).

splistnumpy.ndarray

Represent the list of starting points information formed in the aggregation. list of [ starting point index of current group, sorting values, and number of group elements ]

radiusfloat

The tolerance to control the aggregation. If the distance between the starting point of a group and another data point is less than or equal to the tolerance, the point is allocated to that group. For details, we refer users to [1].

sort_valsnumpy.ndarray

Sorting values.

half_nrm2numpy.ndarray

Precomputed values for distance computation.

Returns

labels_setlist

Connected components of graph with groups as vertices.

connected_pairs_storelist

List for connected group labels.

References

[1] X. Chen and S. Güttel. Fast and explainable sorted based clustering, 2022