August 8, 2005

IBM announces the debut of IBM WebSphere Information Integrator OmniFind Edition, which will provide an open source search alternative to enterprises.

IBM also announced that the UIMA framework itself will be supported by a number of partners, including business intelligence (BI) standouts Cognos and SAS, customer relationship management (CRM) player Kana, and many text analytics providers.

UIMA has been in development for four years or so, and offers a way to get past the simple utility of keyword-based searches, according to analyst Ian Charlesworth of Ovum. "UIMA is designed to look for and interpret the hidden relationships that exist between discrete unstructured data items, exploiting a range of analysis techniques such as natural language processing, machine learning, and automated reasoning."

IBM offered an example of what these hidden relationships could be. For example, there could be product quality information captured in maintenance and repair records, call center logs, and warranty claims that could give you a heads-up on which parts/products are problematic, and why. That, in turn, is actionable data for engineering or other organizations.

This example can be extended into any scenario in which there is a repository of unstructured information from which a business wants to draw broader conclusions about its products or its service.

As Charlesworth points out, making data actionable in this context involves using search that goes beyond keywords and that can get into other techniques.

While he's sanguine about the technology, he points out that there is an enormous counterweight to IBM's presence in this market: "certain parts of the market will be reticent, and will inevitably look to Microsoft to address the issue."

Source: Line 56

Advertise on E-Commerce News