Triangle Count coefficient that are returned are incorrect.

[tomzeppenfeldt]

I have the impression that

CALL algo.triangleCount.stream('Person', 'KNOWS', {concurrency:4})
YIELD nodeId, triangles, coefficient

RETURN algo.asNode(nodeId).id AS name, triangles, coefficient
ORDER BY coefficient DESC

as illustrated on https://neo4j.com/docs/graph-algorithms/current/algorithms/triangle-counting-clustering-coefficient/ (table 6.25) returns the wrong results:

from https://en.wikipedia.org/wiki/Clustering_coefficient :
The local clustering coefficient {\displaystyle C_{i}} C_{i} for a vertex {\displaystyle v_{i}} v_{i} is then given by the proportion of links between the vertices within its neighbourhood divided by the number of links that could possibly exist between them. For a directed graph, {\displaystyle e_{ij}} e_{ij} is distinct from {\displaystyle e_{ji}} e_{{ji}}, and therefore for each neighbourhood {\displaystyle N_{i}} N_{i} there are {\displaystyle k_{i}(k_{i}-1)} k_{i}(k_{i}-1) links that could exist among the vertices within the neighbourhood ( {\displaystyle k_{i}} k_{i} is the number of neighbours of a vertex).

For node “Michael”, 3 out of 6 possible edges connecting all vertices in their neighbourhood actually exist, so the coefficient should be 0,5 instead of 0,3