question about seed file

80 views
Skip to first unread message

masoud....@gmail.com

unread,
Dec 19, 2013, 12:24:05 PM12/19/13
to junto...@googlegroups.com
 Hi,

 I want to use adsorption algorithm on a text file and found junto so useful for this purpose. I almost understood how to use this toolkit; but, still I'm wondering what the seed file is used for. Could you please help me figure out what how to fill the seed file?

thanks

Partha Pratim Talukdar

unread,
Dec 19, 2013, 12:51:46 PM12/19/13
to junto...@googlegroups.com
On Thu, Dec 19, 2013 at 12:24 PM, <masoud....@gmail.com> wrote:

I want to use adsorption algorithm on a text file and found junto so useful for this purpose. I almost understood how to use this toolkit; but, still I'm wondering what the seed file is used for. Could you please help me figure out what how to fill the seed file?

The seed file specifies the label supervision provided to the algorithm. In particular, it stores the initial labels for some subset of nodes in the graph. The algorithm uses these initial seeded nodes, and the underlying graph to label the remaining unlabeled nodes. I hope this helps.

~Partha

laraib...@stonybrook.edu

unread,
Dec 8, 2016, 5:37:02 PM12/8/16
to The Junto Label Propagation Toolkit Open Discussion, partha....@cs.cmu.edu
Quick question about this: what does the number in the seed file represent in the third column? Is it a weight or initial probability for the label? 

Partha Talukdar

unread,
Dec 8, 2016, 11:21:31 PM12/8/16
to junto...@googlegroups.com
On Fri, Dec 9, 2016 at 4:07 AM, <laraib...@stonybrook.edu> wrote:
Quick question about this: what does the number in the seed file represent in the third column? Is it a weight or initial probability for the label? 

It is the initial score for the seed label on the seeded node.

Reply all
Reply to author
Forward
0 new messages