Method and system for generating scored recommendations based on scored references

• US 8,126,898 B2
• Filed: 11/06/2006
• Issued: 02/28/2012
• Est. Priority Date: 11/06/2006
• Status: Active Grant

• Current Assignee: Salesforce.com, Inc.
• Original Assignee: Salesforce.com, Inc.
• Inventors: Constantinides, Paul
• Primary Examiner(s): Lee, Wilson
• Assistant Examiner(s): PHAN, TUANKHANH D
• Application Number:
  US11/593,783
• Publication Number:
  US 20080109426A1
• Time in Patent Office: 1,940 Days
• Field of Search: 707/706, 707/723, 707/748, 707/749, 707/752, 715/838
• US Class Current: 707/748
• CPC Class Codes: G06F 16/3347 using vector based model
• CPC Class Codes: G06Q 10/06 Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
• CPC Class Codes: G06Q 10/10 Office automation; Time management

Abstract

A computer method includes receiving a query from a user system; searching a plurality of data objects of a first type for data objects that substantially match the query; generating a first-relevance score for each data object of the second type that is associated with at least one data objects of the first type identified in the search; searching a plurality of data objects of the second type for data objects of the second type that substantially match the query; generating a second-relevance score for each data object of the second type identified in the second search; generating a list of data objects of the second type that are identified in the first and second search steps; ranking the data objects of the second type in the list of data objects based on the relevance scores; and transferring the list of data objects to the user system.

Claims

1. A computer method comprises:

receiving a query from a user system;

generating a query string that substantially matches characters of the query;

searching a plurality of data objects of a first type for data objects of the first type that substantially match the query wherein i) each data object of the first type is associated with at least one data object of a second type, ii) each data object of the first type includes a query, and iii) each data object of the second type includes an answer to a query;

generating a first-relevance score for each data object of the second type that is associated with at least one of the data objects of the first type that was identified in the searching step;

searching a plurality of data objects of the second type for data objects of the second type that substantially match the query string;

generating a second-relevance score for each data object of the second type that substantially matches the query;

generating a list of data objects of the second type that are identified in the search of the data objects of the first type and that are identified in the search of the data objects of the second type;

ranking the data objects of the second type in the list of data objects based on the first and second relevance scores; and

transferring the list of data objects to the user system.

2. The method of claim 1, wherein the data objects of the first type are case objects.

3. The method of claim 2, wherein the data objects of the second type are solution objects.

4. The method of claim 3, further comprising:

generating a case object for the query;

receiving a request from the user system for the generation of an association the case object for the query and a set of solution objects in the list of data objects; and

forming an association between the case object for the query and the set of solution objects in the list of data objects.

5. The method of claim 4, further comprising providing for subsequent quires the case object for the query and the set of solution objects in the list of data objects.

6. The method of claim 1, wherein the step of generating the first-relevance score for each of the data objects of the second type identified in the first search includes:

i) generating a first-relevance score for each of the data objects of the first type identified in the first search, and ii) assigning the first-relevance score for each of the data objects of the first type to one or more data objects of the second type associated with the data object of the first type.

7. The method of claim 1, wherein the step of generating the second-relevance score for each of the data objects of the second type identified in the second search includes adding a weight to each of the second-relevance scores to generate weighted second-relevance scores (WSRSs).

8. The method of claim 7, wherein each weight is C (relevance score (C)+weighting factor)/count (C).

9. The method of claim 8, wherein:

i) C represents a sum over all data object of the first type associated with the data object of the second type,

ii) the relevance scores (C) are the relevance scores for each of the data objects of the first type associated with the data object of the second type,

iii) the weighting factor is a constant divided by the number of data objects of the first type associated with the data object of the second type, and

iv) the count (C) is a function of the number of data objects of the first type associated with the data object of the second type.

10. The method of claim 9, wherein:

i) if the number of the data objects of the first type associated with the data objects of the second type is less than a low threshold, then the count (C) is the low threshold minus one

ii) if the number of the data objects of the of the first type associated with the data objects of the second type is greater than or equal to the low threshold and is less than or equal to a high threshold, then the count (C) is the number of data objects of the first type associated with the data objects of the second type, and

iii) if the number of the data objects of the first type is greater than the high threshold, then the count (C) is (the high threshold +(the number of data object of the first type associated with the data objects of the second type −

the high threshold)½

).

11. The method of claim 7, wherein the step ranking the data objects of the second type in the list of data objects based on the first and second relevance scores includes ranking the data objects of the second type in the list of objects based on i) the first-relevance score for the data objects of the second type identified in the first search and ii) the second relevance score plus the WSRSs.

12. The method of claim 1, wherein the step of transferring the list of data objects to the user system includes transferring the list of data objects via a web page, via an e-mail, and/or via an instant message.

13. The method of claim 1, further comprising, based on a threshold number, purging a set of data object from the list of data object that is transferred to the user system.

14. A computer method comprising:

receiving a query from a user system;

generating a query string from the query;

searching a plurality of case objects for case objects that substantially match the query string, wherein;

i) at least one of the case objects is associated with at least one solution object,

ii) each of the case objects includes a query, and

iii) each of the solution objects includes an answer to a query;

generating a first-relevance score for each of the solution objects that is associated with at least one of the case objects identified in the searching step;

searching a plurality of solution objects for solution objects that substantially match the query string;

generating a second-relevance score for each of the solution objects that substantially matches the query;

generating a list of solution objects identified in the search of the plurality case objects and identified in the search of the plurality of solution objects;

ranking the solution objects in the list of solution objects based on the first and second-relevance scores; and

transferring the list of solution objects to the user system.

15. The computer method of claim 14, further comprising generating the query string by:

i) removing from the query;

words shorter than a first select length,

words longer than a second select lengths, which is longer than the first select length, and/or

stop words;

ii) stemming words and/or

iii) performing bi-gram tokenization.

16. The method of claim 14, further comprising:

generating a case object for the query;

receiving a request from the user system for the generation of an association the case object for the query and a set of solution objects in the list of solution objects; and

forming an association between the case object for the query and the set of solution objects in the list of solution objects.

17. The method of claim 16, further comprising providing for subsequent quires the case object for the query and the set of solution objects in the list of data objects.

18. A computer system comprising

a server system configured to:

receive a query from a user system;

generate a query string based on the query;

search a plurality of case objects for case objects that substantially match the query string, wherein;

i) at least one of the case objects is associated with at least one solution object,

ii) each of the case objects is associated with includes a query, and

iii) each of the solution objects includes an answer to a query;

generate a first-relevance score for each of the solution objects that is associated with at least one of the case objects identified in the searching step;

search a plurality of solution objects for solution objects that substantially match the query string;

generate a second-relevance score for each of the solution objects that substantially matches the query;

generate a list of solution objects identified in the search of the plurality case objects and identified in the search of the plurality of solution objects;

rank the solution objects in the list of solution objects based on the first and second-relevance scores; and

transfer the list of solution objects to the user system.

19. The computer system of claim 18, wherein the server system is configured to generate the query string by:

1) removing from the query;

words shorter than a first select length,

words longer than a second select lengths, which is longer than the first select length, and/or

stop words;

2) stem words and/or

3) perform bi-gram tokenization.

20. The computer system of claim 18, further comprising a user system communicatively coupled to the server system, wherein the user system is configured to generate the query and transfer the query to the server system, and receive the list of ranked solution objects from the server system.

21. The method of claim 18, further comprising a network configured to i) communicatively couple the user system and the server system, ii) transport the query from the user system to the server system, and iii) transport the list of ranked solution objects from the server system to the user system.