Back to results
US 8,712,934 B2Granted
Temporal document trainer and method
AssigneeJohn Nicholas Gross
InventorJohn Nicholas Gross
Life of this patent
Estimating the term…
Abstract
An electronic document sorter is trained to classify documents based on their temporal qualities. The invention can be used in environments such as automated news aggregators, search engines and other electronic systems which compile information having temporal qualities.
Claims27
A method of automatically training and operating an electronic document sorter with a computing system comprising:
- a) processing a first set of reference electronic documents containing content associated with a first reference story concerning one or more topics and associated events using the computing system;
- b) processing said first set of set of electronic documents for said first reference story to identify a set of events, including identifying from temporal components a temporal relationship of a first event of said first reference story relative in time to a separate second event of said first reference story identified in said set of events;
- wherein said set of events and temporal components are derived using natural language processing and using content other than but not excluding date/time data specified for said first reference story;
- c) training an automated document classifier using said temporal components and said set of events for said first reference story;
- d) operating said automated document classifier to process a second unclassified document containing content for an unclassified story and so as to detect a third event identified for said one or more topics within such unclassified document;
- wherein the automated document classifier is adapted by such training to classify temporal components of said second unclassified electronic document so as to order said third event identified for said one or more topics in said unclassified story in time relative to said first event and/or said second event.
3 more elements
A method of automatically training and operating an electronic document sorter with a computing system comprising:
- a) processing a first set of set of reference electronic documents containing content associated with a first reference story concerning one or more topics and associated events using the computing system;
- b) sorting said first set of reference electronic documents in temporal order with respect to said one or more topics and associated events to form at least one sorted topic/event document set for said first reference story;
- wherein said sorting is performed by the computing system using human-provided annotations and/or input identifying a temporal rank of each of said first set of reference electronic documents;
- c) analyzing said at least one sorted topic/event document set using the computing system to identify temporal content relating to each electronic document which correlates to a temporal rank identified for such electronic document;
- wherein said temporal content is derived using natural language processing and includes content other than but not excluding date/time data specified for said first reference story document;
- d) training an automated document classifier using the computing system so that it can automatically identify such temporal content in an unclassified electronic document containing content relating to said first reference story which is not part of said sorted topic/event document set;
- wherein the automated document classifier is adapted to classify temporal content of such unclassified document to assign a temporal rank of such document relative to other ones of said first set of reference electronic documents for said first reference story.
4 more elements
Dependencies are inferred from the claim text.