This object literal represents a set of optional parameters which can be used in the creation of a ClusteringTransformer.

Since

2016.0

Hierarchy

  • ClusteringParameters

Properties

clusterShapeProvider?: ClusterShapeProvider

An object that defines the location at which the center of the cluster will be placed. By default, the location of the element closest to the center of mass of the contained elements is chosen.

clusterSize?: number

Approximate cluster size in pixels. This size indicates the area of the cluster in screen space. For example, when passing 200 as value, the clustering algorithm will try to create clusters that are 200 pixels apart.

Note that the given size is an approximate size. The clustering algorithm may create smaller or larger clusters, depending on the data. The size of the cluster is influenced by:

  • This setting
  • The proximity of other clusters. Different clusters can start competing, trying to include the same point. This will cause clusters to be closer together than indicated by this setting, and thus be smaller.
  • The amount of points in the cluster. Clusters with a small amount of points will break down sooner than clusters with many points. This means that clusters with a small amount of points will typically be smaller.
minimumPoints?: number

Minimum number of elements with the same classification required to create a cluster. By default, a cluster must contain at least 2 objects.

noClustering?: boolean

Indicates whether clustering should be applied. By default, clustering is enabled. If noClustering is set to true, the other parameters won't be taken into account as no data will be clustered.