Minutes from Non 12 Telcon

[11:56] <Ashok> meeting:Property Graphs CG Nov 12, 2013
[11:57] <Ashok> chair: Ashok
[12:05] <Ashok> present: Michael, Ashok, Zhe, Patrick, Gregg
[12:05] <Ashok> present+: Danny
[12:06] <Ashok> Minutes from Last weekhttp://lists.w3.org/Archives/Public/public-propertygraphs/2013Nov/0004.html
[12:06] <zwu2> looks good
[12:06] <Ashok> Approved without objection
[12:06] <Ashok> Topic: New members
[12:08] <Ashok> Danny:Graphlab is an open source project for machine learning ... 4 years old ... several users
[12:08] <Ashok> ... developing tools and toolkits
[12:09] <Ashok> ... successful in collaborative filtering
[12:09] <Ashok> See graphlab.com
[12:10] <Ashok> Patrick:ISO Stds development ... topic maps, graphs and hypergraphs
[12:11] <Ashok> ... more theoretical, construction that could be represented as property graphs
[12:12] <Ashok> Topic:Usecases
[12:15] <Ashok> Zhe:http://lists.w3.org/Archives/Public/public-propertygraphs/2013Nov/0007.html
[12:16] <Ashok> Overview:In a society, an individual's behavior can be heavily influenced by that individual's close friends.Examples of such behaviors include but not limited to alcohol consumption, health related habits,like/dislike products, and adopting/rejecting services. Specifically in the telecommunication domain,service providers pay a lot of attention to spotting potential churners.
To the service providers, keepingexisting customers is as important as gaining new customers. One way to identify potential churners is to1) model the underlying social network and call network as a graph with attributes describing users andtheir interactions, 2) run graph analytics against the graph to identify communities,and 3) rank existing users based on the attrition rate of their close friends in a community, andpossibly other signals.
[12:19] <Ashok> Zhe:Churn is a big problem in telecom area ... need to identify churners
[12:19] <Ashok> ... partition graph into smaller graphs
[12:20] <Ashok> ... graph clustering, vertex ranking
[12:20] <Ashok> ... many algorithms can be used for clustering
[12:21] <Ashok> ... eliminate weak links
[12:22] <pdurusau> Q+
[12:23] <Ashok> Patrick:Is a good usecase ... because Property Graphs can have properties on the arcs
they can indicate the nature of relationships
[12:24] <@mpetyx> Q+
[12:25] <Ashok> Michael:I'm fine with this
------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
[12:27] <zwu2> q+
[12:27] <Ashok> Danny, Patrick ... can contribute usecases
[12:28] <zwu2> Danny, you mentioned a few other popular tool kits, what are those?
[12:28: <Ashok> Danny may have left the call
[12:29] <zwu2> thanks & bye
-- 
All the best, Ashok