Text Analytics in Search (and a 'PS' on Inxight)

The relationship between search and text analytics was a recurring topic at this week's Text Analytics Summit in Boston. The one supports information retrieval and the other just about anything else automated you can do with a document set, from knowledge extraction to automated classification and processing: complementary functions that rely on similar technical underpinnings. Ramana Rao, who has a wonderful ability to clarify, put it this way: "Google's white box makes everything seem so simple," but "we got to simplicity without handing the complexity of reality." It's text analytics, of course, that will equip search to handle complexity.I believe that the pervasiveness of search is a symptom of a failure of design, of presumptuous attempts to structure information access that hinder findability as often as they boost it. David Bean, co-founder and CTO of Attensity, clearly also thinks along these lines. "Attensity is making war on knowledge engineering. We want to take the human out of the loop."

Search is an enabling technology toward that end, but search is not enough. According to Hadar Shemtov of Yahoo, for major Web-search providers, "text mining is everywhere in a sense." Text mining fuels semantically enriched search indexes and powers delivery of contextual advertising, which is of course the service that generates the bulk of his company's and Google's revenues. "Assisted search and query rewriting" allow the engines to provide results that reflect what the user really intends, which would otherwise be difficult to discern given the terseness of searches that average under three words in length. These various forms of text analytics enable emerging next-generation search, "active information supply driven by user intention and by context."

Summit speakers and panelists identified many unmet text-analytics challenges. The challenges ranged from boosting NLP and classification accuracy to creation of scalable, embedded text technologies to ROI considerations that are key to wider implementation. Regardless, the variety and quality of the presentations, and of the tools and applications described, demonstrated one point quite well. Text-analytics is delivering serious value, both to direct end-users and, especially, for the wide world of search users - for everyone who uses the Web.

P.S. A source, not now or formerly from either company, told me that my estimate of the price Business Objects is paying for Inxight, however well rationalized it might have been, was way high. He claimed to know definitively that the Inxight sale price is $72 million.

Seth Grimes is president of consultancy Alta Plana Corporation and chair of the Text Analytics Summit.The relationship between search and text analytics was a recurring topic at this week's Text Analytics Summit in Boston. The one supports information retrieval and the other just about anything else automated you can do with a document set: complementary functions that rely on similar technical underpinnings. Ramana Rao, who has a wonderful ability to clarify, put it this way: "Google's white box makes everything seem so simple," but "we got to simplicity without handing the complexity of reality."