“Considering the glut of information on the Web today, categorizing documents by hand is no longer time- or cost-effective, nor is it consistent,” said Ian Hersey, Inxight’s vice president of linguistic products. “News agencies, Web portal builders, and other online publishers depend on high-speed coding to deliver accurate material. Categorizer is an ideal tool that combines significant flexibility with the high quality of a manual editorial review. Corporate IT organizations can also greatly benefit from using Categorizer on their internal portals to reduce the effort required to maintain their intranets.”
Categorizer uses Inxight’s patented linguistic analysis technologies to compare and classify new documents with a collection of manually coded documents and classifies them into categories. Categorizer differs from rule-based classification systems because it learns how to identify topics in a document from previously coded examples. In addition, the Categorizer assigns a confidence value to each code it attaches to a document based on its ability to identify similar training documents.
Categorizer software is adaptable to different writing styles, is scalable to over 100,000 documents per day with a single server, and uses an Oracle-based data store for high-availability environments. It also provides true linguistic support for Western European languages and optional support for company coding using Inxight Thing Finder.
Inxight Software was established in 1996 by the Xerox Corp. as part of the Xerox New Enterprise business initiative. Inxight’s mission is to capitalize on breakthrough user-interface, linguistic, and information-access technologies invented and developed at Xerox’s Palo Alto Research Center (PARC) and Xerox Research Center Europe (XRCE) in Grenoble, France. Other Inxight products include Summarizer (for text mining); Thing Finder (for entity extraction); and several data-visualization products, including Hyperbolic Tree, Site Lens Studio, and Table Lens. Additional product information on Categorizer is available at http://www.inxight.com/products/enterprise/categorizer.html.
Source: Inxight Software, Inc., Palo Alto, CA, 877/INXIGHT; http://www.inxight.com.
|Table of Contents||Information Today Home Page|