Abstract: Embodiments disclose a method of obtaining relevancy scores for concepts found on a page of content. Methods disclosed rely on connectivity of concepts found in a page in relation to connectivity of the same concepts with other concepts in general, and the semantic relationships between the concepts found on a page to obtain relevancy of concepts.
Type:
Application
Filed:
March 2, 2011
Publication date:
September 8, 2011
Applicant:
Knewco, Inc.
Inventors:
Erik Van Mulligen, Ravi Kalaputapu, Marc Weeber
Abstract: Data structures, systems, methods and computer program products that enable precise information retrieval and extraction, and thus facilitate relational and associative discovery are disclosed. The present invention utilizes a novel data structure termed a “Knowlet” which combines multiple attributes and values for relationships between concepts. While texts contain many re-iterations of factual statements, Knowlets record relationships between two concepts only once and the attributes and values of the relationships change based on multiple instances of factual statements, increasing co-occurrence or associations. The present invention's approach results in a minimal growth of the Knowlet space as compared to the text space and it thus useful where there is a vast data store, a relevant ontology/thesaurus, and a need for knowledge navigation and (relational, associative, and/or other) knowledge discovery.
Type:
Grant
Filed:
March 31, 2008
Date of Patent:
August 2, 2011
Assignee:
Knewco, Inc.
Inventors:
Albert Mons, Nickolas Barris, Christine Chichester, Barend Mons, Erik van Mulligen, Marc Weeber