News   Features   White Papers   Research Reports   Web Events   Source Book   Conferences  
Entity extraction meets federated search
Posted Nov 21, 2005 Print Version     Page 1of 1

Inxight ( has unveiled SmartDiscovery Awareness Server 5.4, which allows users to query a variety of internal and external sources from one search box and then cluster and filter the results by the most relevant people, companies, concepts (noun phrases), places and other entities mentioned in them, speeding access to information.

For example, the company says, government analysts can access open and classified sources with a single query and then see and filter by the most relevant concepts the weapons, organizations and people mentioned in documents to pinpoint information of interest. Further, it adds, brand managers can search multiple news agencies, analyst reports and internal information to see the companies, concepts and products most often mentioned in conjunction with their own company and a search term of interest. Citing another example, drug researchers can access Medline, subscription journals and other relevant Web sites to see and filter results by the genes, kinases and other information most commonly associated with a given search term.

The server is preconfigured out of the box to search and track relevant data from more than 600 online and enterprise sources. It ranks this information by relevance and provides a variety of intelligent filtering and preview options, including entity, category and source clustering. Customized alerts and page tracking automatically inform users of new and updated information, explains Inxight.

Entity extraction is accomplished when used with Inxight SmartDiscovery Analysis Server, which, the company says, reads text and automatically detects people, companies, places, concepts (noun phrases) and many other entities out of the box, without the need for complicated taxonomies, tuning or pre-defined lists. Users can also create their own list-based and pattern-based entities for use in filtering.

Print Version   Page 1of 1

Problems with this site? Please contact the webmaster. | About ITI | Privacy Policy