Adding CURE clustering algorithm to WEKA

Question

I have written a java program to perform CURE clustering. I wish to add this program to weka as a clustering algorithm and visualize the clustering. Has anyone already implemented it on weka?Any links to that would be very much helpful. How do I proceed with it?

http://weka.sourceforge.net/doc.dev/weka/clusterers/package-summary.html - This is the JavaDoc for the clustering functionality in WEKA. This seems to me a good place to start reading if you wish to implement new clusterer in WEKA. Also check the corresponding source codes. — zloster, Nov 18 '11 at 20:50

score 3 · Answer 1 · answered Nov 25 '11 at 17:29

Weka is very weak on clustering. Actually clustering in Weka is pretty much non-existant. The implementations of e.g. OPTICS are essentially external programs that are just called, but not at all integrated. You might want to have a look at ELKI, http://elki.dbs.ifi.lmu.de/ which is like Weka for clustering and outlier detection (it doesn't have classification though - this is where Weka is strong). It does not yet have CURE though, either. But a dozen of other clusterers and visualizations for them.

score 2 · Answer 2 · answered Apr 13 '12 at 15:28

2

There is a CURE on Weka project available here: http://code.google.com/p/data-mining-cure-algorithm

answered Apr 13 '12 at 15:28

Jon Reindeer

23
4

Adding CURE clustering algorithm to WEKA

2 Answers2