CBOD(label: String = "community", cutoff: Double = 0.0, labeler:GraphAlgorithm = Identity())

Returns outliers detected based on the community structure of the Graph.

The algorithm assumes that the state of each vertex contains a community label (e.g., set by running LPA on the graph, initially) and then defines an outlier score based on a node’s community membership and how it compares to its neighbors community memberships.


label: String = "community"

Identifier for community label (default: “community”)

cutoff: Double = 0.0

Outlier score threshold (default: 0.0). Identifies the outliers with an outlier score > cutoff.

labeler: GraphAlgorithm

Community algorithm to run to get labels (does nothing by default, i.e., labels should be already set on the input graph, either via chaining or defined as properties of the data)


outlierscore: Double

Community-based outlier score for vertex


(only for vertex such that outlierscore >= cutoff)

vertex name

outlier score

name: String

outlierscore: Double

See also