[
https://issues.apache.org/jira/browse/SPARK-13733?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15207219#comment-15207219
]
Gayathri Murali commented on SPARK-13733:
-----------------------------------------
[~mengxr] [~dwmclary] https://issues.apache.org/jira/browse/SPARK-5854 -
mentions one key difference between page rank and personalized page rank as:
"In PageRank, every node has an initial score of 1, whereas for Personalized
PageRank, only source node has a score of 1 and others have a score of 0 at the
beginning.", which basically means we initialize as [1, 0, 0, 0, ..] where we
set 1 only to the seed node.
For this JIRA, do we want to instead set an initial distribution of weights
such that all nodes receive non-zero initial values? Could you clarify if this
is the behavior that is intended?
> Support initial weight distribution in personalized PageRank
> ------------------------------------------------------------
>
> Key: SPARK-13733
> URL: https://issues.apache.org/jira/browse/SPARK-13733
> Project: Spark
> Issue Type: New Feature
> Components: GraphX
> Reporter: Xiangrui Meng
>
> It would be nice to support personalized PageRank with an initial weight
> distribution besides a single vertex. It should be easy to modify the current
> implementation to add this support.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
