HOLISTIC DISAMBIGUATION FOR ENTITY NAME SPOTTING

• US 20100223292A1
• Filed: 02/27/2009
• Published: 09/02/2010
• Est. Priority Date: 02/27/2009
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Grandison, Tyrone W.A., Gruhl, Daniel F., Pieper, Jan H., Bhagwan, Varun
• Granted Patent: US 8,856,119 B2
• Time in Patent Office: Days
• US Class Current: 707/780
• CPC Class Codes: G06F 40/295 Named entity recognition

Abstract

A method resolves ambiguous spotted entity names in a data corpus by determining an activation level value for each of a plurality of nodes corresponding to a single ambiguous entity name. The activation levels for each of the nodes may be modified by inputting outside domain knowledge corresponding to the nodes to increase the activation value of the nodes, spotting entity names corresponding to the nodes to increase the activation value of the nodes, searching the data corpus to spot newly posted entity names to increase the activation value of the nodes, and searching the data corpus to reduce or deactivate the activation value of the nodes by eliminating false positives. The ambiguous entity name is assigned to the node determined to have the highest activation level and is then outputted to a user.

Claims

1. A computer implemented method for reducing ambiguities in entity name spotting, comprising:

performing an entity name spotting process in a data corpus;

identifying, based on said entity name spotting process, an ambiguous entity name, said ambiguous entity name comprising an entity name corresponding to at least two potentially matching categorical nodes of an activation network, said activation network comprising a plurality of predefined categorical nodes;

determining an activation level for each of said potentially matching categorical nodes;

selecting a most activated categorical node of said potentially matching categorical nodes, said most activated categorical node having a highest activation level of each potentially matching categorical node; and

outputting said most activated categorical node to a user to replace said ambiguous entity name.

2. The method of claim 1, wherein, for each categorical node, said determining of said activation level comprises:

receiving metadata related to entity names in said data corpus to increase said activation level of said potentially matching categorical nodes;

analyzing said metadata to determine related categorical nodes to increase said activation level of said potentially matching categorical nodes;

analyzing semantic information of user comments to increase said activation level of said potentially matching categorical nodes; and

analyzing an ontology customized for said activation network to increase said activation level of said potentially matching categorical nodes,

wherein said entity name comprises one of proper names, credentials and identifications.

3. The method of claim 2, wherein said metadata is from one of said data corpus and a subject matter expert data source.

4. The method of claim 2, wherein said augmentation semantic information comprises one of a domain specific forum and a domain specific blog corresponding to said activation network.

5. The method of claim 2, wherein said analyzing of said ontology determines at least one of an acronym and a nickname of said categorical nodes.

6. A computer implemented method for reducing ambiguities in entity name spotting, comprising:

creating one node for each entity name of interest of a data corpus comprising a plurality of entity names, each of said nodes having an activation level;

inputting outside domain knowledge to change activation levels of said nodes;

searching a user forum to spot entity names;

modifying said activation levels on nodes corresponding to entity names spotted in said forum during said searching;

searching said user forum for newly posted entity names;

additionally modifying activation levels on nodes corresponding to said newly posted entity names;

identifying ambiguous entity names from said data corpus, said ambiguous entity names each corresponding to a plurality of potentially matching nodes;

selecting a most activated node from said potentially matching nodes, said most active node having a highest activation level;

assigning said ambiguous entity name to said most activated node;

once all ambiguous entity names have been assigned, continue searching said data corpus for next newly posted entity names; and

outputting said most activated categorical node to a user to replace said ambiguous entity name,

wherein said entity name comprises one of proper names, credentials and identifications.

7. The method of claim 6, wherein said inputting outside domain knowledge comprises a subject matter expert data source.

8. The method of claim 6, wherein said modifying said activation level utilizes semantic information comprising one of a domain specific forum and a domain specific blog.

9. The method of claim 6, wherein said modifying said activation level utilizes ontology to determine one of an acronym and a nickname corresponding to said entity name.

10. The method of claim 6, wherein said modifying said activation level comprises one of reduction and deactivation of said activation level by eliminating false positives utilizing at least one of stopwords and heuristics.

11. A computer implemented method for reducing entity name spotting ambiguities in an activation network comprising a plurality of nodes representing predefined categories of items of interest within a domain, each of said nodes comprising an activation level, said method comprising:

gathering domain specific text from a plurality of wide area network accessible sources to provide a data corpus, said data corpus comprising a plurality of entity names;

searching said data corpus to spot entity names, said entity names corresponding to at least one node of said plurality of nodes;

modifying said activation level of said nodes based on entity names spotted during said searching;

identifying an ambiguous entity name, said ambiguous entity name corresponding to a plurality of potentially matching nodes;

selecting a most activated node of said potentially matching nodes, said most activated node having a highest activation level of each potentially matching node; and

outputting said most activated categorical node to a user to replace said ambiguous entity name,

wherein said entity name comprises one of proper names, credentials and identifications.

12. The method of claim 11, wherein said modifying said activation level comprises retrieving static information derived from information sources corresponding to said spotted entity name.

13. The method of claim 11, wherein said modifying said activation level comprises semantically analyzing domain specific user comments corresponding to said spotted entity name.

14. The method of claim 11, wherein said modifying said activation level comprises ontologically analyzing said entity names for one of acronyms and nicknames.

15. The method of claim 11, wherein said modifying said activation level comprises one of reduction and deactivation of said activation level by eliminating false positives using at least one of stopwords and heuristics.

16. A computer implemented method for reducing ambiguities in entity name spotting within an activation network, comprising:

creating one node for each entity name of interest of a data corpus comprising a plurality of entity names, each of said nodes having an activation level;

modifying a value of said activation level of said nodes by one of;

inputting outside domain knowledge corresponding to said nodes to increase said activation value of said nodes;

searching said data corpus to spot entity names corresponding to said nodes to increase said activation value of said nodes; and

searching said data corpus to spot newly posted entity names to increase said activation value of said nodes; and

searching said data corpus to one of reduce and deactivate said activation value of said nodes by eliminating false positives;

identifying ambiguous entity names from said data corpus, said ambiguous entity names each corresponding to a plurality of potentially matching nodes;

assigning said ambiguous entity name to said most activated node; and

outputting said most activated categorical node to a user to replace said ambiguous entity name,

wherein said entity name comprises one of proper names, credentials and identifications.

17. The method of claim 16, further comprising:

once all ambiguous entity names have been assigned, continue searching said data corpus for next newly posted entity names.

18. The method of claim 16, wherein said outside domain knowledge comprises a subject matter expert data source.

19. The method of claim 16, wherein said data corpus comprises one of a domain specific forum and a domain specific blog corresponding to said activation network.

20. The method of claim 16, wherein said searching said data corpus to one of reduce and deactivate said activation value comprises using at least one of stopwords and heuristics.