DBSCANConfig

Properties

eps

number> 0

Radius of neighborhood.

min_points

integer≥ 1

Minimum number of points in a neighborhood to form a cluster.

weights

array | null≥ 1 items

Weights for data points. NULL = unweighted. Must have one value per case.

border_points

boolean

Assign border points to clusters.

search

string

Nearest neighbor search strategy.

one of"kdtree""linear""dist"

bucket_size

integer≥ 1

Size of buckets for the k-d tree search.

split_rule

string

Rule for splitting the k-d tree.

one of"SUGGEST""STD""MIDPT""FAIR""SL_MIDPT""SL_FAIR"

approx

boolean

Use approximate nearest neighbor search.

Relationships