Method and system for performing an authority analysis

摘要

Methods and systems for automatically determining, from a body of emails, blogs, and other documents, authors of the documents who are authorities on certain subjects, and what those subjects are. An intersection of the semantic footprints of documents by an author are deemed to be the derived skills footprint of the author. The derived skills footprints of many authors are compared with a user's query to determine who is the best person that could respond to the user.

主权项

1. A method of recommending an authority on a subject, the method comprising:
analyzing a plurality of documents for each of a plurality of authors to determine their semantic footprints; assigning a rating to each of the plurality of documents based upon the respective semantic footprints and usage data associated with each of the plurality of documents; determining, using a processor communicatively coupled to a memory, an author of highly rated documents having a intersecting semantic footprint to establish a subject of authority for the author, the processor determining which documents are highly rated documents based upon the usage data associated with each of the plurality of documents; associating the intersecting semantic footprint with the author; receiving a query from a user; analyzing the query for its semantic footprint; determining when the semantic footprint of the query intersects the intersecting semantic footprint associated with the author; determining the extent of the intersection between the semantic footprint of the query and the intersecting semantic footprint associated with the author based upon a ratio of a union of the semantic footprint of the query and the intersecting semantic footprint associated with the author; and providing, when the semantic footprint of the query intersects the intersecting semantic footprint associated with the author, the user with contact information for the author, wherein each semantic footprint is a multi-dimensional coordinate based on one or more keywords, entities and annotations, and wherein each semantic footprint is determined based upon a volume of overlapping areas of the respective multi-dimensional coordinates.