Method and apparatus for concept-based ranking of natural language discourse

• US 9,063,970 B1
• Filed: 11/01/2011
• Issued: 06/23/2015
• Est. Priority Date: 06/10/2005
• Status: Active Grant

• Current Assignee: Netbase Solutions, Inc.
• Original Assignee: Netbase Solutions, Inc.
• Inventors: Rehling, John Andrew, Osofsky, Michael Jacob
• Primary Examiner(s): NGUYEN, KIM T
• Application Number:
  US13/286,799
• Time in Patent Office: 1,330 Days
• Field of Search: 707/705, 707/706
• US Class Current: 1/1
• CPC Class Codes: G06F 16/24 Querying
• CPC Class Codes: G06F 16/243 Natural language query formulation
• CPC Class Codes: G06F 16/24578 using ranking
• CPC Class Codes: G06F 16/3344 using natural language analysis

Abstract

Concept-based ranking is applied to the results of a search process. Concept-based ranking relies on the input query, that produced the search results, having pinnacle concept properties specified for some of its keyword-based search terms. For each record “r” of a search result, it can be assigned a concept-based ranking score as follows. For each query term “QT,” of the input query, the following can be performed. The set of occurrences of “QT,” in the record, can be determined. Based upon the pinnacle concept of “QT,” and certain characteristics of the set of occurrences, a value can be assigned to “QT.” Characteristics, of the set of occurrences, can include the type of pinnacle concept asserted by at least one sub-unit of the record in which there is an occurrence. Once each “QT” has been assigned a value, such values can be combined to produce a score for ranking the record.

Claims

1. A method for assigning a score for ranking a record, given the record and a query, comprising:

identifying, performed at least in part with a configuration of computing hardware and programmable memory, a first query term, of the query;

identifying, performed at least in part with a configuration of computing hardware and programmable memory, a first pinnacle concept property of the first query term;

identifying, performed at least in part with a configuration of computing hardware and programmable memory, a first set of occurrences of the first query term in the record;

identifying, performed at least in part with a configuration of computing hardware and programmable memory, a first set of sub-units of the record, wherein each sub-unit contains a member of the first set of occurrences;

determining, performed at least in part with a configuration of computing hardware and programmable memory, for a first sub-unit of the first set of sub-units, whether a first concept, referenced by the first sub-unit, is a pinnacle concept, and identification of the first concept as a pinnacle concept indicates that the first concept is not a plains concept;

assigning, performed at least in part with a configuration of computing hardware and programmable memory, a set of pinnacle concepts to each member of the first set of sub-units, wherein a pinnacle concept is a member, of a set of pinnacle concepts, if the pinnacle concept is asserted by the sub-unit, of the first set of sub-units, to which it is assigned;

determining whether the first pinnacle concept property specifies a first pinnacle concept;

assigning, if the first pinnacle concept is specified, a first score to the first query term that is larger when the first pinnacle concept is a member of a set of pinnacle concepts of the first set of sub-units than when the first pinnacle concept is not a member of a set of pinnacle concepts of the first set of sub-units;

assigning, if the first pinnacle concept is not specified, a first score to the first query term that is larger when the first set of sub-units is not empty than when the first set of sub-units is empty;

assigning, performed at least in part with a configuration of computing hardware and programmable memory, a first score to the first query term as a function of the first pinnacle concept property and first set of sub-units; and

assigning, performed at least in part with a configuration of computing hardware and programmable memory, a score, for ranking the record, as a function of the first score.

2. The method of claim 1, further comprising:

determining that the first pinnacle concept property specifies a first pinnacle concept;

determining that the first pinnacle concept is not a member of a set of pinnacle concepts of the first set of sub-units; and

assigning a first score to the first query term that is larger when at least one set of pinnacle concepts, for a sub-unit of the first set of sub-units, is empty of a correlative pinnacle concept of the first pinnacle concept than when at least one set of pinnacle concepts, for a sub-unit of the first set of sub-units, contains a correlative pinnacle concept of the first pinnacle concept.

3. The method of claim 1, further comprising:

identifying a second query term, of the query;

identifying a second pinnacle concept property of the second query term;

identifying a second set of occurrences of the first query term in the record;

identifying a second set of sub-units of the record, wherein each sub-unit contains a member of the second set of occurrences;

assigning a set of pinnacle concepts to each member of the second set of sub-units, wherein a pinnacle concept is a member, of a set of pinnacle concepts, if the pinnacle concept is asserted by the sub-unit, of the second set of sub-units, to which it is assigned;

assigning a second score to the second query term as a function of the second pinnacle concept property and second set of sub-units; and

assigning a score, for ranking the record, as a function of the first and second scores.

4. The method of claim 3, wherein the function of the first and second scores determines a product of the first and second scores.

5. The method of claim 1, further comprising:

applying a classification procedure to the record to determine a set of pinnacle concepts referenced by sub-units of the record.

6. The method of claim 1, further comprising:

searching a first database, according to the query, to produce a first search result of which the record is a part;

assigning a score, for ranking a record, to records of the first search result other than the record; and

ranking records of the first search result according to the score for ranking a record.

7. The method of claim 1, wherein the record represents a snippet.

8. The method of claim 1, wherein a sub-unit is a sentence.

9. A system, for assigning a score for ranking a record, given the record and a query, comprising:

a sub-system configured, as a result of its computing hardware and programmable memory, to accomplish identifying a first query term, of the query;

a sub-system configured, as a result of its computing hardware and programmable memory, to accomplish identifying a first pinnacle concept property of the first query term;

a sub-system configured, as a result of its computing hardware and programmable memory, to accomplish identifying a first set of occurrences of the first query term in the record;

a sub-system configured, as a result of its computing hardware and programmable memory, to accomplish identifying a first set of sub-units of the record, wherein each sub-unit contains a member of the first set of occurrences;

a sub-system configured, as a result of its computing hardware and programmable memory, to accomplish determining, for a first sub-unit of the first set of sub-units, whether a first concept, referenced by the first sub-unit, is a pinnacle concept, and identification of the first concept as a pinnacle concept indicates that the first concept is not a plains concept;

a sub-system configured, as a result of its computing hardware and programmable memory, to accomplish assigning a set of pinnacle concepts to each member of the first set of sub-units, wherein a pinnacle concept is a member, of a set of pinnacle concepts, if the pinnacle concept is asserted by the sub-unit, of the first set of sub-units, to which it is assigned;

a sub system configured to accomplish determining, of whether the first pinnacle concept property specifies a first pinnacle concept;

a sub system configured to accomplish assigning, if the first pinnacle concept is specified, of a first score to the first query term that is larger when the first pinnacle concept is a member of a set of pinnacle concepts of the first set of sub-units than when the first pinnacle concept is not a member of a set of pinnacle concepts of the first set of sub-units;

a sub system configured to accomplish assigning, if the first pinnacle concept is not specified, of a first score to the first query term that is larger when the first set of sub-units is not empty than when the first set of sub-units is empty;

a sub-system configured, as a result of its computing hardware and programmable memory, to accomplish assigning a first score to the first query term as a function of the first pinnacle concept property and first set of sub-units; and

a sub-system configured, as a result of its computing hardware and programmable memory, to accomplish assigning a score, for ranking the record, as a function of the first score.

10. The system of claim 9, further comprising:

a sub system configured to accomplish determining, that the first pinnacle concept property specifies a first pinnacle concept;

a sub system configured to accomplish determining, that the first pinnacle concept is not a member of a set of pinnacle concepts of the first set of sub-units; and

a sub system configured to accomplish assigning, of a first score to the first query term that is larger when at least one set of pinnacle concepts, for a sub-unit of the first set of sub-units, is empty of a correlative pinnacle concept of the first pinnacle concept than when at least one set of pinnacle concepts, for a sub-unit of the first set of sub-units, contains a correlative pinnacle concept of the first pinnacle concept.

11. The system of claim 9, further comprising:

a sub system configured to accomplish identifying, of a second query term of the query;

a sub system configured to accomplish identifying, of a second pinnacle concept property of the second query term;

a sub system configured to accomplish identifying, of a second set of occurrences of the first query term in the record;

a sub system configured to accomplish identifying, of a second set of sub-units of the record, wherein each sub-unit contains a member of the second set of occurrences;

a sub system configured to accomplish assigning, of a set of pinnacle concepts to each member of the second set of sub-units, wherein a pinnacle concept is a member, of a set of pinnacle concepts, if the pinnacle concept is asserted by the sub-unit, of the second set of sub-units, to which it is assigned;

a sub system configured to accomplish assigning, of a second score to the second query term as a function of the second pinnacle concept property and second set of sub-units; and

a sub system configured to accomplish assigning a score, for ranking the record, as a function of the first and second scores.

12. The system of claim 11, wherein the function of the first and second scores determines a product of the first and second scores.

13. The system of claim 9, further comprising:

a sub system configured to accomplish applying, of a classification procedure to the record to determine a set of pinnacle concepts referenced by sub-units of the record.

14. The system of claim 9, further comprising:

a sub system configured to accomplish searching of a first database, according to the query, to produce a first search result of which the record is a part;

a sub system configured to accomplish assigning a score, for ranking a record, to records of the first search result other than the record; and

a sub system configured to accomplish ranking, of records of the first search result according to the score for ranking a record.

15. The system of claim 9, wherein the record represents a snippet.

16. The system of claim 9, wherein a sub-unit is a sentence.