Inferring relations between internet objects that are not connected directly

• US 7,299,270 B2
• Filed: 07/10/2001
• Issued: 11/20/2007
• Est. Priority Date: 07/10/2001
• Status: Expired due to Term

• Current Assignee: RPX Corporation
• Original Assignee: Lycos Incorporated (Brightcom Group Ltd.)
• Inventors: Kotlov, Andrei, DoRosario, Alden, Guidi, John, Balasubramaniam, Shan, Kolluri, Venkateswarlu
• Primary Examiner(s): Cardone, Jason
• Assistant Examiner(s): Bhatia, Ajay M
• Application Number:
  US09/902,421
• Publication Number:
  US 20030101286A1
• Time in Patent Office: 2,324 Days
• Field of Search: 707/104.1, 707 2- 7, 709/218, 719/316, 715/513
• US Class Current: 709/218
• CPC Class Codes: G06F 16/9558 Details of hyperlinks; Management of linked annotations
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

An inferred relation weighting process for determining the strength of an inferred relation between a first and a second Internet object which are not directly linked includes a first link weighting process for determining the strength of at least a first link between the first non-directly linked Internet object and a common object. A second link weighting process determines the strength of at least a second link between the second non-directly linked Internet object and the common object. An inferred relation weight calculation process calculates the strength of the inferred relation based on the strength of the at least a first link and the at least a second link.

Claims

1. A computer-implemented inferred relation weighting process for determining a strength of an inferred relation between a first internet object and a second Internet object, where the first and second Internet objects are not directly linked, comprising:

a first link weighting process for determining a first strength of a first link between the first Internet object and a common object;

a second link weighting process for determining a second strength of a second link between the second Internet object and the common object, the first and second link weighting processes being performed by one or more processors; and

an inferred relation weight calculation process for determining the strength of inferred relation based on the first strength and the second strength;

wherein the first Internet object comprises a query for retrieving a document and the second Internet object comprises a document, the inferred relation weight calculation process comprising;

determining a first array having entries comprising the first and second strengths;

determining a second array as a mathematical function of the first array, wherein the first and second arrays are stored in memory;

determining a weighted sum of entries selected from the first and second arrays, the weighted sum corresponding to the strength of the inferred relation; and

returning a group of Internet objects associated with the first Internet object to a user, the group including the second Internet object, the second Internet object being arranged within the group according to the strength of the inferred relation.

2. The inferred relation weighting process of claim 1, wherein the common object comprises a plurality of Internet objects, the plurality of Internet objects being interconnected via discrete links, the plurality of Internet objects being connected to the first and second links;

and

wherein said inferred relation weighting process further comprises;

an intermediate link weighting process for determining a strength of each discrete link, wherein the strength of the inferred relation is based on a strength of each discrete link.

3. The inferred relation weighting process of claim 1, wherein the common object comprises at least one Internet document.

4. The inferred relation weighting process of claim 2, further comprising:

a link limitation process for specifying a link limit concerning a maximum number of links used to determine the inferred relation.

5. The inferred relation weighting process of claim 2, further comprising:

an incoming link analysis process for determining a number of objects linked to each Internet object, wherein an incoming link value of a target Internet object is proportional to a number of objects linked to the target Internet object.

6. The inferred relation weighting process of claim 2, further comprising:

an outgoing link analysis process for determining a number of objects that each Internet object is linked to, wherein an outgoing link value of a target Internet object is proportional to a number of objects to which the target Internet object is linked.

7. The inferred relation weighting process of claim 2, wherein the inferred relation weight calculation process comprises a relation recalculation process for redefining values corresponding to strengths of discrete links and to the first and second strengths in response to calculation of the strength of the inferred relation.

8. The inferred relation weighting process of claim 1, wherein at least one of the Internet objects is a transaction record.

9. The inferred relation weighting process of claim 1, wherein at least one of the Internet objects comprises an Internet document.

10. The inferred relation weighting process of claim 1, wherein the strength of the inferred relation corresponds to a relevance score.

11. The inferred relation weighting process of claim 8, wherein the relevance score comprises a percentage.

12. A computer-implemented inferred relation weighting process for determining a strength of an inferred relation between a first Internet object and a second Internet object, where the first and second Internet objects are not directly linked, comprising:

a first link weighting process for determining a first strength of a first link between the first Internet object and a plurality of common objects;

a second link weighting process for determining a second strength of a second link between the second Internet object and the plurality of common objects, wherein the plurality of common objects comprises a first common object connected to the first link, a second common object connected to the second link, and an intermediate link between the first and second common objects;

an intermediate link weighting process for determining a strength of the intermediate link, the first, second, and intermediate link weighting processes being performed by one or more processors;

an inferred relation weight calculation process for determining the strength of the inferred relation based on the first strength, the second strength, and the strength of the intermediate link;

wherein the first Internet object comprises a query for retrieving a document and the second Internet object comprises a document, the inferred relation weight calculation process comprising;

determining a first array having entries comprising the first and second strengths;

determining a second array as a mathematical function of the first array, wherein the first and second arrays are stored in memory;

determining a weighted sum of entries selected from the first and second arrays, the weighted sum corresponding to the strength of the inferred relation; and

returning a group of Internet objects associated with the first Internet object, the group including the second Internet object, the second Internet object being arranged within the group according to the strength of the inferred relation.

13. The inferred relation weighting process of claim 12, further comprising:

a link limitation process for specifying a link limit concerning a maximum number of links used to determine the inferred relation.

14. The inferred relation weighting process of claim 12, wherein the plurality of common objects comprises at least one Internet document.

15. The inferred relation weighting process of claim 12, wherein the intermediate link comprises at least one additional common object and a plurality of sub-links for connecting the at least one additional common object to the first and second common objects;

and

wherein the intermediate link weighting process determines the strength of the intermediate link based on the individual strengths of the sub-links.

16. The inferred relation weighting process of claim 15, further comprising:

an incoming link analysis process for determining a number of objects linked to each Internet object and common object, wherein an incoming link value of a target object is proportional to number of objects linked to the target object.

17. The inferred relation weighting process of claim 15, further comprising:

an outgoing link analysis process for determining a number of objects that each Internet object and common object is linked to, wherein an outgoing link value of a target object is proportional to a number of objects to which the target object is linked.

18. The inferred relation weighting process of claim 15, wherein the inferred relation weight calculation process comprises a relation recalculation process for redefining values corresponding to strengths of the sub-links and to the first and second strengths in response to calculation of the strength of the inferred relation.

19. The inferred relation weighting process of claim 12, wherein at least one of the Internet objects comprises a transaction record.

20. The inferred relation weighting process of claim 12, wherein at least one of the Internet objects comprise an Internet document.

21. The inferred relation weighting process of claim 12, wherein the strength of the inferred relation corresponds to a relevance score.

22. The inferred relation weighting process of claim 21, wherein the relevance score comprises a percentage.

23. A computer-implemented method for determining a strength of an inferred relation between a first Internet object and a second Internet object, where the first and second Internet objects are not directly linked, the method comprising:

determining a first strength of a first link between the first Internet object and a common object;

determining a second strength of a second link between the second Internet object and the common object;

determining a strength of the inferred relation based on the first strength and the second strength, wherein the first strength, the second strength, and the strength of the inferred relation are determined by one or more processors;

wherein the first Internet object comprises a query for retrieving a document and the second Internet object comprises a document, and wherein calculating a strength of the inferred relation comprises;

determining a first array having entries comprising the first and second strengths;

determining a second array as a mathematical function of the first array; and

determining a weighted sum of entries selected from the first and second arrays,

wherein the first and second arrays are stored in memory, the weighted sum corresponding to the strength of the inferred relation; and

returning a group of Internet objects associated with the first Internet object, the group including the second Internet object, the second Internet object being arranged within the group according to the strength of the inferred relation.

24. The method of claim 23, wherein the common object comprises a plurality of Internet objects, the plurality of Internet objects being connected via discrete links, wherein determining the strength of the inferred relation further comprises:

determining a strength of each discrete link, wherein the strength of the inferred relation is based also on a strength of each discrete link.

25. The method of claim 24, further comprising:

specifying a link limit concerning a maximum number of links used to determine the inferred relation.

26. A machine-readable medium for storing instructions for implementing an inferred relation weighting process, the inferred relation weighting process for determining a strength of an inferred relation between a first Internet object and a second Internet object, where the first and second Internet objects are not directly linked, wherein the instructions, when executed by a processor, cause the processor to:

determine a first strength of a first link between the first Internet object and a common object;

determine a second strength of a second link between the second Internet object and the common object;

determine the strength of the inferred relation based on the first strength and the second strength by;

determining a first array having entries comprising the first and second strengths;

determining a second array as a mathematical function of the first array; and

determining a weighted sum of entries selected from the first and second arrays, the weighted sum corresponding to the strength of the inferred relation;

wherein the first Internet object comprises a query for retrieving a document and the second Internet object comprises a document; and

return a group of Internet objects associated with the first Internet object, the group including the second Internet object, the second Internet object being arranged within the group according to the strength of the inferred relation.

27. The machine-readable medium of claim 26, wherein the machine-readable medium comprises a random access memory (RAM).

28. The machine-readable medium of claim 26, wherein the machine-readable medium comprises a read only memory (ROM).

29. The machine-readable medium of claim 26, wherein the machine-readable medium comprises a hard disk drive.

30. An apparatus for executing an inferred relation weighting process for determining a strength of an inferred relation between a first Internet object and a second Internet object, where the first and second Internet objects are not directly linked, the apparatus comprising:

memory to store instructions that are executable; and

at least one processing device to execute the instructions to;

determine a first strength of a first link between the first Internet object and a common object;

determine a second strength of a second link between the second Internet object and the common object; and

determine the strength of the inferred relation based on the first strength and the second strength by;

determining a first array having entries comprising the first and second strengths;

determining a second array as a mathematical function of the first array;

determining a weighted sum of entries selected from the first and second arrays, the weighted sum corresponding to the strength of the inferred relation;

wherein the first Internet object comprises a query for retrieving a document and the second Internet object comprises a document; and

returning a group of Internet objects associated with the first Internet object, the group including the second Internet object, the second Internet object being arranged within the group according to the strength of the inferred relation.

31. The apparatus of claim 30, wherein the processor and memory are part of into a personal computer.

32. The apparatus of claim 30, wherein the processor and memory are part of into a network server.

33. The apparatus of claim 30, wherein the processor and memory are part into a single board computer.