Raytheon BBN Technologies

Transform data into usable information.

Quickly sift through documents, web pages, and email to discover relevant information. IdentiFinder Text Suite™ solves the classic problems of text mining: first, how to identify significant documents and then, how to locate the most important information within them.

Bridging the Gap: IdentiFinder Text Suite™

Bridge the gap between unstructured text and structured data stores with BBN IdentiFinder Text Suite™, a software tool that rapidly analyzes electronically-stored text to locate names of corporations, organizations, people, and places, including variations in names. IdentiFinder uses top-performing patented statistical algorithms that not only spot these “named entities” within text, but also identify the types of names, distinguishing, for example between Virginia the state and Virginia a person.

Let IdentiFinder Text Suite help you manage your information flow by:

• Grouping names into aliases (“IBM” and “Big Blue” for International Business Machines)
• Spotting names on watch lists as well as new names that are impossible to specify in advance
• Selecting documents for translation
• Searching, sorting, and indexing documents
• Learning new entities as they occur
• Locating entities regardless of formatting such as HTML, all capitals, mixed case, or lowercase.

Out of the box, IdentiFinder Text Suite™ recognizes up to 24 types of entities, such as:

• Persons
• Dates & Times
• Organizations
• Money
• Locations
• ...and more!

IdentiFinder Delivers Where Rule-based Systems Fail

IdentiFinder’s statistical methods can extract entities in a wide range of text sources.

Competitors’ rule-based systems work fine as long as target data matches the rules, but fail when faced with new data sources, varied document structures and content, and noisy, ungrammatical data. In addition to handling well-written news wire copy, IdentiFinder can locate misspelled entities and extract from instant messaging, blogs, and other less formal text sources.

Highly scalable and fault tolerant, IdentiFinder Text Suite’s robust extraction is ready out of the box to start solving your problems and degrades gracefully in the face of noisy input.

Train IdentiFinder to Improve Extraction Performance

With the Learner Module you can improve performance without writing tedious and fragile rules. Missing names, wrong extents, incorrect names—no matter what the problem, BBN IdentiFinder is both easier and faster to customize than any other solution. Users simply provide examples and the software learns on its own. No training necessary! Our competition’s systems can only be improved by experts trained to carefully craft rules.