Is it possible to use Apache Mahout for term extraction? The term can be words or phrases.
That's a broad question. I don't think its purpose as a project directly includes term extraction, though the TF-IDF implementation probably has a lot of the pieces and building blocks that might be of use in making such an implementation.