[ https://issues.apache.org/jira/browse/FLINK3780?page=com.atlassian.jira.plugin.system.issuetabpanels:commenttabpanel&focusedCommentId=15291192#comment15291192
]
ASF GitHub Bot commented on FLINK3780:

Github user vasia commented on a diff in the pull request:
https://github.com/apache/flink/pull/1980#discussion_r63890500
 Diff: docs/apis/batch/libs/gelly.md 
@@ 2051,6 +2052,26 @@ The algorithm takes a directed, vertex (and possibly edge) attributed
graph as i
vertex represents a group of vertices and each edge represents a group of edges from
the input graph. Furthermore, each
vertex and edge in the output graph stores the common group value and the number of represented
elements.
+### Jaccard Index
+
+#### Overview
+The Jaccard Index measures the similarity between vertex neighborhoods. Scores range
from 0.0 (no common neighbors) to
+1.0 (all neighbors are common).
+
+#### Details
+Counting common neighbors for pairs of vertices is equivalent to counting the twopaths
consisting of two edges
 End diff 
By "twopaths" you mean triads? i.e. open triangles?
> Jaccard Similarity
> 
>
> Key: FLINK3780
> URL: https://issues.apache.org/jira/browse/FLINK3780
> Project: Flink
> Issue Type: New Feature
> Components: Gelly
> Affects Versions: 1.1.0
> Reporter: Greg Hogan
> Assignee: Greg Hogan
> Fix For: 1.1.0
>
>
> Implement a Jaccard Similarity algorithm computing all nonzero similarity scores. This
algorithm is similar to {{TriangleListing}} but instead of joining twopaths against an edge
list we count twopaths.
> {{flinkgellyexamples}} currently has {{JaccardSimilarityMeasure}} which relies on {{Graph.getTriplets()}}
so only computes similarity scores for neighbors but not neighborsofneighbors.
> This algorithm is easily modified for other similarity scores such as AdamicAdar similarity
where the sum of endpoint degrees is replaced by the degree of the middle vertex.

This message was sent by Atlassian JIRA
(v6.3.4#6332)
