I'm organizing a bakeoff, if you want to show off some Mahout skills
and do a controlled comparison of Mahout to other people's approaches:
Let's say I have several hundred million documents, which are very
short (only a few words). There are several million terms in the
vocabulary. What is the fastest way to find the top-k semantically
related terms for each term in the vocabulary?
If you just want to hear the results, join this group:
http://groups.google.com/group/metaoptimize-challenge-announce
If you actually want to hack some data, read this blog post:
http://metaoptimize.com/blog/2010/11/05/nlp-challenge-find-semantically-related-terms-over-a-large-vocabulary-1m/
It would be really cool to see participation from the Mahout community
in a Mahout demo, to get a controlled comparison to other
implementations.
Best,
Joseph
|