Hello jyoti, 

I guess for k map clustering the first centroids you choose is most of the time 
random though you can make good intuitive selection that would reduce the 
number of later iterations. And then after getting the first set of cluster (by 
assigning each node to the centroid it is closest to), in the next iteration 
you change the centroid and make the one that is closest to the Mean of the 
cluster. 

Based on this, the node running the clustering algorithm should change. Of 
course that's what you do during the k-mean clustering. 

I hope this helps. 

Regards,
Rizwan Mumtaz

 

On 10/20/11, jyoti patole  <[email protected]> wrote:

> Hi
> 
> I m working on "Clustering in Wireless sensor netwotk"
> & i m implementing K means with map reduce  algorithm. using Tiny OS
> 
> I  need to know , at which node the clustering algorithm must be executed?
> 
> 
> Is it Base station, Cluster Head, or any special node must be used?
> 
> Please reply
> 
> -- 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> With Warm Regards
> 
> JYOTI
> 
> 9766319445
> 
> 
> 
> _______________________________________________
> Tinyos-help mailing list
> [email protected]
> https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

_______________________________________________
Tinyos-help mailing list
[email protected]
https://www.millennium.berkeley.edu/cgi-bin/mailman/listinfo/tinyos-help

Reply via email to