Text mining for the rest of us

Fellow Corante Web Hub contributor, Matthew Hurst wants Consumer Facing Text Mining Opportunities.  He wants text mining for non-commercial customers that actually provides some value to the user.

The basic technologies required to do this: NLP (information extraction, parsing, semantic interpretation, summarization), data mining, text mining are all being deployed in enterprise facing solutions. They are being applied to a certain extent in consumer solutions, but their results are never surfaced in those solutions as the primary result.

This clearly connects to the wide world of aggregation, social filtering, smart aggregators and the like.  I may not have the right language to talk about the text mining features, but I suspect there is some help to be found from these kinds of tools.

I would like to echo what Marshall Kirkpatrick commented on my recent Smartening the aggregator post that I don't want automatic filtering and grouping on everything.  I want the tool to give me the option of saying, "please find other stuff like this," where I have many options on what "like this" means. 

I could see a lot of value in using a text mining-enabled tool to help me do ongoing research where I don't necessarily want to read the river of news in the given area.  For example, I'd love an agent to bring me articles that mention "knowledge management," that aren't in my regular suite of reading, and that come from a trusted sources.  Oh, and provide me a useful summary of what's come up since I last looked.  It's those last bits that are difficult, particularly for a fairly high-level topic like knowledge management.  At this point, I have many web feeds for searches on the term in blogs and beyond.  But when the results come back, I have to flip through them manually to figure out if they are of any value.  And of course, those which only provide excerpts don't always give me enough information.