On Sun, 20 Sep 2009 15:58:05 -0700 (PDT) jakobitsch juergen <[email protected]> wrote:
> - could someone point me into the right direction for basic intro > into keyphrase extraction using mahout So far I am not aware of people using Mahout for key-phrase extraction. I am not myself familiar with kea but know of people on-list having experience with kea. Would be interesting to read their comments as well. There are a few publications available online that deal with training a classifier to identify key-phrases. Basically the general approach would be to first manually label a set of "good" and "bad" phrases. Next one would have to come up with features describing these phrases and training a classifier on this example set. Mahout does come with a set of classification algorithms, but has no full-featured end-to-end solution for the problem of key-phrase identification. You would still need to manually label example phrases, come up with a good feature set and integrate the Mahout training algorithms into your software. Isabel
