From discussions and Q&A with attendees at the Predictive Analytics World conference - I can personally confirm manya's first projection. Namely when she said 2009 challenges will call for
"A broader set of vertical/horizontal offerings including more automated unstructured (text, voice, image) capabilities delivered for customer/product/competitive intelligence"Although text mining was not addressed in the title or absract of any of the 30 talks on the agenda there were 3 separate discussions that did include ways text analytics fit in. My favorite was when John Elder applied text analytics to "build his haystack" out of the pile of Social Security disability claim forms. By applying text mining, 20% of the cases could be automatically approved quickly leaving more time for the manual review of the others. John's point was that while Text Mining didn't specifically find the needle in the haystack , the technology most certainly helped organize and arrange the haystack so he could allocate tasks to the "artifical intelligence" of the computer and help people do their job faster.
Yes , I say 2009 will be the year business and industries open their eyes to see untapped potential for insight now laying around as unstructured content...for example see this blog where Jason Burke says
As pharmaceutical firms have learned, you can create great solutions for sharing data, and the information still not be useful. For example, if data is sitting in an EMR as an unstructured block of free text, your ability to get to better decisions and predictable outcomes will be considerably constrained. Thankfully, we do have more mature interoperability standards today that can help us along the way. But interoperability should not just be about pushing data across systems -- it should be about facilitating medical insight across stakeholders
Jason has laid the groundwork ...and i wait with baited breath to see his next blog entry - because its the INSIGHT that excites me about analytics.
Text analytics is not promising easy answers - infact it may spur additional new questions for investigation --resulting in worthwhile innovation. Finding case studies will be a challenge for 2009 - just as the yahoo group discussion here has pointed out.