[
https://issues.apache.org/jira/browse/MAHOUT-344?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12849598#action_12849598
]
Ankur commented on MAHOUT-344:
------------------------------
Appreciate your interest in this. I'd suggest that we pick a dataset to try
this on and then make changes as required.
For ideas :-
1. Clustering a p2p dataset like this -
http://warsteiner.db.cs.cmu.edu/db-site/Datasets/graphData/eDoneky-p2p/ to find
out nodes closer to each other.
2. Clustering similar items/songs in this -
http://www.iua.upf.es/~ocelma/MusicRecommendationDataset/index.html for
recommendations.
Talking about missing things in the implementation:-
1. Option of more hash functions for user's to experiment with.
2. Code for cluster goodness evaluation (Precision/Recall tests?)
3. Unit tests for completeness.
May be other Mahout folks can take a quick look at the patch and suggest more
ideas.
> Minhash based clustering
> -------------------------
>
> Key: MAHOUT-344
> URL: https://issues.apache.org/jira/browse/MAHOUT-344
> Project: Mahout
> Issue Type: Bug
> Components: Clustering
> Affects Versions: 0.3
> Reporter: Ankur
> Assignee: Ankur
> Attachments: MAHOUT-344-v1.patch
>
>
> Minhash clustering performs probabilistic dimension reduction of high
> dimensional data. The essence of the technique is to hash each item using
> multiple independent hash functions such that the probability of collision of
> similar items is higher. Multiple such hash tables can then be constructed
> to answer near neighbor type of queries efficiently.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
