Available Technology

A root-and-rule-based approach to information indexing and searching

The invention is a method and apparatus to index collections of documents through a combination of natural language processing tools and the application of a novel rules-and-root-based method for automatically creating phrase-based, structured keywords that encode semantics for facilitating searches. The system provides the ability to search for terms and their known context of appearance (semantics) to disambiguate the intended search results. The apparatus is designed for continuous incremental evolution of terminology in any domain. The output of this method also serves as a starting point for creating domain-based taxonomy of concepts.

Abstract: 

The invention is a method and apparatus to index collections of documents through a combination of natural language processing tools and the application of a novel rules-and-root-based method for automatically creating phrase-based, structured keywords that encode semantics for facilitating searches. The system provides the ability to search for terms and their known context of appearance (semantics) to disambiguate the intended search results. The apparatus is designed for continuous incremental evolution of terminology in any domain. The output of this method also serves as a starting point for creating domain-based taxonomy of concepts.

Inventors: 

T. N. Bhat, J. Elliott, U. Kattner, C. Campbell, R. Siram, E. Subrahmanian, J. Collard, I. Monarch

Patent Number: 
15/883449
Internal Laboratory Ref #: 
17-007
Patent Status: 
pending application 15/883,449
Patent Issue Date: 
June 25, 2019
Lab Representatives
Share to Facebook Share to Twitter Share to Google Plus Share to Linkedin