Hello I am trying to implement label propagation algorithm for community detection. I have some questions to clear confusions about it
My scenario is I have large data set having different labels of nodes. I have loaded those from CSV files. Each node label hase some different property. as Example I have a label of customer nodes. where the properties are "CustomerID","location","division","Regdate". on the other hand the relationship consists properties like AMOUNT,CHANNEL, TXNID, TIMESTAMP
-
could I use a node property as seed label ? or do I need to generate random value as seed label and update it as distinct property name seed label. the documentation said that It should be a numeric value. but most numeric values are residing in relationship properties
-
could I use label propagation algorithm for different node labels. like I have customer label and merchant label. where
(n:Customer)-[r:PAYMENT]->(m:Merchant)
could I project subgraph for this , I tried it in same label In customer-[relations]-customer through documentation. but couldn't solve it with different label of nodes.
kindly help me to clear this concept
