System and method for processing a query

• US 20050080780A1
• Filed: 08/20/2004
• Published: 04/14/2005
• Est. Priority Date: 08/21/2003
• Status: Active Grant

• Current Assignee: Idilia, Inc.
• Original Assignee: Idilia, Inc.
• Inventors: Colledge, Matthew, Carrier, Marc
• Granted Patent: US 7,509,313 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 40/247 Thesauruses; Synonyms
• CPC Class Codes: G06F 40/30 Semantic analysis
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99934 Query formulation, input preparation, or translation
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

The invention provides a system and method of processing a query directed to a database. The invention comprises implementing the steps of: obtaining the query from a user; and disambiguating the query using a knowledge base to obtain a set of identifiable senses associated with words in the query. Further if the set comprises more than one identifiable sense, then the following additional steps are executed: selecting one sense from the set as a best sense; utilizing the best sense of the query to identify relevant results from the database related to the best sense; re-disambiguating the remaining senses of the set by excluding results associated with the best sense; selecting a next best sense from the remaining senses; and utilizing the next best sense of the query to identify relevant results from the database related to the next best sense. The invention also provides updates to databases for users, sessions and common data relating to the best identified results for the queries, to improve and personalize disambiguation of subsequent queries by a user.

Claims

1. A method of processing a query directed to a database, said method comprising the steps of:

obtaining said query from a user;

disambiguating said query using a knowledge base to obtain a set of interpretations for words in said query;

if said set comprises more than one identifiable interpretation

selecting one interpretation from said set as a best interpretation;

utilizing said best interpretation of said query to identify relevant results from said database related to said best interpretation;

re-disambiguating remaining interpretations of said set by excluding results associated with said best interpretation;

selecting a next best interpretation from said remaining interpretations; and

utilizing said next best interpretation of said query to identify relevant results from said database related to said next best interpretation.

2. The method of processing a query directed to a database as claimed in claim 1, wherein said step of disambiguating said query comprises utilizing an algorithm selected from:

an example disambiguation algorithm, an n-word disambiguation algorithm, a priors disambiguation algorithm;

a dependencies algorithm and a classifying algorithm.

3. The method of processing a query directed to a database as claimed in claim 2, further comprising:

obtaining a selected identifiable sense from said user identifying a selected word sense from said set of identifiable senses; and

updating said knowledge base with data model regarding said query and said selected identifiable sense.

4. The method of processing a query directed to a database as claimed in claim 3, wherein said data model comprises data model for said user.

5. The method of processing a query directed to a database as claimed in claim 4, wherein said data model further comprises data for a session associated with said user and said query.

6. The method of processing a query directed to a database as claimed in claim 5, wherein said data model further comprises data model for all queries to said database.

7. A method of processing a query directed to a database, said method comprising the steps of:

obtaining said query from a user;

disambiguating said query using a knowledge base to obtain a set of identifiable interpretations associated with words in said query;

if said set comprises more than one identifiable interpretation;

selecting one sense from said set as a best interpretation;

utilizing said best interpretation of said query to identify relevant results from said database related to said best interpretation;

expanding said best interpretation to obtain related word senses for said best interpretation to produce an expanded best interpretation of said query;

compare said expanded best interpretation of said query to an index associated with said database;

selectively processing remaining interpretations of said set by;

re-disambiguating said remaining interpretations of said set by excluding results associated with said best interpretation;

selecting a next best interpretation from said remaining senses;

utilizing said next best interpretation of said query to identify relevant results from said database related to said next best interpretation; and

identifying a term associated with said next best interpretation which distinguishes said next best sense from said best interpretation;

obtaining results from said database utilizing said expanded best interpretation of said query;

generating a question to said user utilizing said term to test whether said next best interpretation was the interpretation meant by said user;

obtaining from said user a response to said question to identify an intended interpretation from said; and

utilizing said response to further re-disambiguate said set of interpretations.

8. The method of processing a query directed to a database as claimed in claim 7, wherein said step of disambiguating said query comprises utilizing an algorithm selected from:

an example disambiguation algorithm, an n-word disambiguation algorithm, a priors disambiguation algorithm, a classifier algorithm and a dependencies algorithm.

9. A method of processing a query directed to a database, said method comprising the steps of:

obtaining said query from a user;

disambiguating said query using a knowledge base to obtain a set of identifiable senses associated with words in said query;

if said set comprises more than one identifiable sense,

selecting one sense from said set as a best sense;

selectively processing remaining senses of said set by;

re-disambiguating said remaining senses of said set by excluding results associated with said best sense; and

selecting at least a next best sense from said remaining senses to form a set of re-disambiguated remaining senses;

for said best sense and each member of said set of re-disambiguated remaining senses;

expanding and paraphrasing its associated sense to obtain semantically related senses to produce an expanded sense of for its query;

comparing its expanded sense to an index associated with said database; and

obtaining results from said database utilizing its expanded sense;

obtaining from said user an indication noting which result from all results returned from said database is an intended sense for said query; and

utilizing said indication to further re-disambiguate said set of senses.

10. A method of revising a knowledge base associated with a query directed to a database, said method comprising the steps of:

disambiguating said query using said knowledge base to obtain a set of identifiable senses associated with words in said query;

identifying an intended sense from said set; and

updating said knowledge base with data regarding said query and said selected identifiable sense to contain preferred identifiable senses which are used to disambiguate subsequent queries.

11. The method of processing a query directed to a database as claimed in claim 10, wherein

said step of disambiguating said query comprises utilizing an algorithm selected from:

an example disambiguation algorithm, an n-word disambiguation algorithm, and a priors disambiguation algorithm; and

said updating said knowledge base involves updating a local knowledge base associated with said algorithm.

12. The method of processing a query directed to a database as claimed in claim 11, wherein said data comprises data for a user.

13. The method of processing a query directed to a database as claimed in claim 12, wherein said data further comprises data for a session associated with said user and said query.

14. The method of processing a query directed to a database as claimed in claim 13, wherein said data further comprises data for a common session of all users accessing said database.