Attensity today launched the beta version of Attensity Server 4.0, the fourth major version of its server application for extracting facts and relationships from unstructured text through its patented suite of extraction engines. Attensity Server 4.0 offers a wizard-based interface for quick, point-and-click extraction of facts from free-form text in millions of documents. It enables seamless import and configuration into structured relational tables in combination with the structured data already stored in databases and accessed by business intelligence applications.

Attensity Server 4.0 provides a single, scalable architecture for enterprise use across multiple computers. This beta version precedes the official release of the Attensity 4.0 product family, designed to provide the next-generation Text Analytics application for business users. It will offer business analysts -- in customer support, marketing, claims processing and government roles -- the ability to search, analyze, query, chart and graph text dynamically through a browser-based, easy-to-use interface.

"Attensity Server 4.0 provides the scalability enterprises require to extract and analyze facts from millions of documents," said Craig D. Norris, Attensity's chief executive officer. "Through a simple, clear, wizard-based interface, users are able to parse vast amounts of unstructured text efficiently and effectively -- without knowledge engineering degrees or heavy involvement from IT professionals."

The management of unstructured data is a large and growing problem. Merrill Lynch has estimated more than 85 percent of all business information exists as unstructured data -- everything from customer emails to service notes to surveillance reports. According to IDC's October 2005 report "Text Mining: Mining for Gold in Unstructured Information" (#CA1503SWD), the worldwide market for natural language understanding software products, which includes Text Analytics, is estimated to reach $1.84 billion by 2008.

With Attensity Server 4.0's new wizard-based, point-and-click extraction process, business users and analysts can easily import their chosen text and extract the "who, what, where, when and why." It also leverages Attensity's patented Exhaustive ExtractionT engine, which automatically and rapidly extracts unstructured data into rows and columns for analysis.

Key features of Attensity Server 4.0 include:

-- Standards: UIMA-compliant wrappers available for all extraction engines. UIMA, or Unstructured Information Management Architecture, is an open source framework published by IBM to promote a standard for connecting text analytics applications that process unstructured information. Also, a new web services (SOAP) layer enables improved automatic access to the extraction engines for integration into other applications.

-- Scheduling: Seamless management of text import and extraction runs, as well as simultaneous processing of text, offers better throughput and data flow management.

-- Intelligence: Industry and customer learning is automatically incorporated into the extraction logic engines for deeper fact and event extraction. Moreover, deeper anaphora resolution offers improved understanding of the common use of pronouns in speech. For example: "Sharon hired Sally. She will begin working for her tomorrow." Attensity Server 4.0 automatically resolves that "she" is Sally and "her" is Sharon.

-- Scalability: Scalability: The application runs on 64-bit Linux Red Hat Enterprise Service 4.0. Distributed servers for any supported platform can be added and removed to scale according to need.

The beta version of Attensity Server 4.0 is available directly from Attensity, as are the recently released Attensity Analytics 2.0 and Attensity Discover 3.0 applications. Additionally, Teradata or IBM sales representatives can discuss comprehensive analytic solutions that integrate Attensity's technology into data management and business intelligence systems.