Semio debuts "text mining" suite

Intranet-based "text mining" software maker Semio rolled out its latest package today.

SemioMap 2.0 is a text mining product that identifies, groups, and maps key concepts within huge stores of unstructured data, and provides users with relevant information in their research and analysis tasks.

The product has a Java applet client and a server-based engine that runs on NT, Solaris, and Unix. Semio is targeting customers in the research and analysis fields.

"This product will be useful anyplace you have a large amount of text research going on, like research reports, financials, reviews of books, pharmaceuticals," Robert Craig, an analyst with the Hurwitz Group, said. "The whole idea of looking for common phrases rather than straight indexing like an Alta Vista solution is very interesting. I was struck by how they look for common relationships between whole trunks of information."

"Text mining" systems are similar to data mining ones in that they deal with large volumes of data and fall under the knowledge discovery niche in the knowledge management market space.

"There are a lot of tools to help knowledge workers find and access data," said Semio's founder and CEO Claude Vogel. "The problem is that there is too much information overload. SemioMap addresses the need to strategically leverage large volumes of text-based data to gain a competitive advantage in today's marketplace."

The underlying technologies of SemioMap combine lexical, or verbal, processing, information clustering, and graphical display. The Semiolex technology, which the product is based on, allows SemioMap to extract key phrases and identify relevant relationships in unstructured information, according to the company.

Prices for the product start at $5,000 per server. It starts shipping today, the company said.