The Data Mining Forum 
IMPORTANT: This is the old Data Mining forum.
I keep it online so that you can read the old messages.
Please post your new messages in the new forum: https://forum2.philippe-fournier-viger.com/index.php
ROCK Clustering Algorithm
Posted by:
ANN
Date: November 13, 2014 01:36AM
]Hi Philippe,
Sorry for bother you. I am just stuck on the goodness measure of rock algorithm formula, I do understand all the process but I could not understand what ni,nj means and what they are represent. I have attached the formula and hope you can help me understand what they are really means. I have read many articles yet they did not explains what they mean and how to get their value to calculate the goodness . so can you help please. Thanks in advance
Sorry for bother you. I am just stuck on the goodness measure of rock algorithm formula, I do understand all the process but I could not understand what ni,nj means and what they are represent. I have attached the formula and hope you can help me understand what they are really means. I have read many articles yet they did not explains what they mean and how to get their value to calculate the goodness . so can you help please. Thanks in advance
Re: ROCK Clustering Algorithm
Posted by:
webmasterphilfv
Date: November 13, 2014 04:46AM
Hi,
Sorry, I'm not familiar with this algorithm. To help you, I would need to also read the paper. Perhaps that someone else on this forum may help you. Otherwise, you may also contact the authors of that paper.
Best,
Sorry, I'm not familiar with this algorithm. To help you, I would need to also read the paper. Perhaps that someone else on this forum may help you. Otherwise, you may also contact the authors of that paper.
Best,
Re: ROCK Clustering Algorithm
Posted by:
arshboy
Date: November 21, 2014 10:20PM
Hie i also have the same problem..Can someone help me too?I already contact them via email but still waiting for their reply for more than 1 month.I really hope that someone could help me.Your help is much appreciated.
Re: ROCK Clustering Algorithm
Posted by:
ameame
Date: December 13, 2014 01:04AM
"As a result, it follows
that if two fairly large clusters with sizes
ni and nj are merged, the number of links between pairs
of points in the merged cluster is (ni+nj)1+2f()"
http://www.cis.upenn.edu/~sudipto/mypapers/categorical.pdf - pg.11
So, these are the sizes of clusters which you merge and then measure the goodness of merging with this formula.
that if two fairly large clusters with sizes
ni and nj are merged, the number of links between pairs
of points in the merged cluster is (ni+nj)1+2f()"
http://www.cis.upenn.edu/~sudipto/mypapers/categorical.pdf - pg.11
So, these are the sizes of clusters which you merge and then measure the goodness of merging with this formula.