Yang Yang created SPARK-10758:
---------------------------------
Summary: approximation algorithms to speedup triangle count and
clustering coefficient computation in GraphX
Key: SPARK-10758
URL: https://issues.apache.org/jira/browse/SPARK-10758
Project: Spark
Issue Type: New Feature
Components: GraphX
Reporter: Yang Yang
We propose to implement an algorithm to exactly count global clustering
coefficient of a given graph, and two approximation algorithms to speedup the
computation of triangle counting and clustering coefficient in GraphX.
The basic idea of the approximation algorithm is: given a large-scale graph, we
sample a subgraph, which is representative in the sense that graph properties
of interest (e.g., #triangles) can be estimated with a known degree of
accuracy.
Our algorithm has been well tested on 16 network data comparing with several
baseline methods (please see details in
https://aminer.org/billboard/network-sampling). For example, when counting
triangles on a Twitter network data (with 2,420,766 edges), the approximation
algorithm can speedup 10x with only 0.36% error rate comparing with exact
algorithm.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
