[
https://issues.apache.org/jira/browse/MAHOUT-153?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12832380#action_12832380
]
Rohini Uppuluri commented on MAHOUT-153:
----------------------------------------
Hi all,
I have implemented an extension to the algorithm Pallavi had mentioned.
The extension is to add some randomness in picking the farthest point. With this
there is a chance of over coming the problem of picking some noise points as
centroids which are very far away.
Summary:
1. Pick the first centroid randomly
2. for the rest of the centroids
do
-> compute a few candidate centroids which are far off
Candidate centroid computation:
Divide the data into few parts.
For each part compute the point which is farthest from
the current list of centroids
-> Select one of the candidate centroids randomly
done
I will soon submit a patch on this. Please let me know your feeback.
> Implement kmeans++ for initial cluster selection in kmeans
> ----------------------------------------------------------
>
> Key: MAHOUT-153
> URL: https://issues.apache.org/jira/browse/MAHOUT-153
> Project: Mahout
> Issue Type: New Feature
> Components: Clustering
> Affects Versions: 0.2
> Environment: OS Independent
> Reporter: Panagiotis Papadimitriou
> Assignee: Ted Dunning
> Fix For: 0.4
>
> Attachments: Mahout-153.patch
>
> Original Estimate: 336h
> Remaining Estimate: 336h
>
> The current implementation of k-means includes the following algorithms for
> initial cluster selection (seed selection): 1) random selection of k points,
> 2) use of canopy clusters.
> I plan to implement k-means++. The details of the algorithm are available
> here: http://www.stanford.edu/~darthur/kMeansPlusPlus.pdf.
> Design Outline: I will create an abstract class SeedGenerator and a subclass
> KMeansPlusPlusSeedGenerator. The existing class RandomSeedGenerator will
> become a subclass of SeedGenerator.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
