System and method for constructing targeted ranking from multiple information sources

• US 8,417,694 B2
• Filed: 08/20/2008
• Issued: 04/09/2013
• Est. Priority Date: 03/31/2008
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Bhagwan, Varun, Grandison, Tyrone Wilberforce Andre, Gruhl, Daniel Frederick, Pieper, Jan Hendrik
• Primary Examiner(s): Beausoliel, Jr., Robert
• Assistant Examiner(s): Rayyan, Susan F
• Application Number:
  US12/195,098
• Publication Number:
  US 20090248614A1
• Time in Patent Office: 1,693 Days
• Field of Search: 707/706, 707/722, 707/723, 707/748
• US Class Current: 707/723
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

Embodiments of the invention provide a system and method for determining preferences from information mashups and, in particular, a system and method for constructing a ranked list from multiple sources. In an exemplary embodiment, the system and method tunably combines multiple ranked lists by computing a score for each item within the list, wherein the score is a function of the associated rank of the item within the list. In one exemplary embodiment, the function is equal to 1/(n^(1/p)), where p is a tuning parameter that enables selection between responsiveness in the combined ranking to one candidate ranked highly in one source versus responsiveness in the combined ranking to a candidate with lower but broader support among the various sources ranking the candidates.

Claims

1. A computer-implemented method for combining multiple ranked lists, the method comprising:

receiving multiple lists, wherein each list comprises items each having an associated rank, n, within the list and wherein at least one of the items is present in more than one of the multiple lists; and

combining the items in the multiple lists into a combined list, wherein the combining comprises;

assigning a value to a variable p, the value being neither 0 nor negative;

for each list in the multiple lists, computing a score for each item within the list, wherein the score is a function of the associated rank within the list, and wherein the function is equal to 1/(n^(1/p));

for each item, summing corresponding scores from the multiple lists for a same item to produce a total score for the item;

ranking each item based on the corresponding total score of each item; and

outputting the combined list based on the ranking.

2. The computer-implemented method of claim 1, wherein assigning a value to a variable p comprises assigning a value of 2.5 to the variable p.

3. The computer-implemented method of claim 2, further comprising:

increasing the value of the variable p to increase ranking in the combined list of items appearing in a majority of the multiple lists; and

repeating the combining.

4. The computer-implemented method of claim 1, further comprising:

decreasing the value of the variable p to increase ranking in the combined list of items having higher than average associated ranking in a minority of the multiple lists; and

repeating the combining.

5. The computer-implemented method of claim 1, further comprising:

receiving a new value for the variable p; and

repeating the combining using the new value as the value for the variable p.

6. The computer-implemented method of claim 1, further comprising:

assigning a new value to the variable p to adjust the ranking in the combined list of items; and

repeating the combining.

7. The computer-implemented method of claim 1, wherein assigning the value to the variable p comprises automatically assigning the value to the variable p based on a predetermined business objective.

8. The computer-implemented method of claim 1, further comprising:

calculating the associated rank n for each item based on a position of the item with the corresponding list in the multiple lists.

9. A computer program product for combining multiple ranked lists, said computer program product comprising:

a non-transitory computer readable medium;

first program instructions to receive multiple lists, wherein each list comprises items each having an associated rank, n, within the list and wherein at least one of the items is present in more than one of the multiple lists; and

second program instructions to combine the items in the multiple lists into a combined list, wherein the combining comprises;

assigning a value to a variable p, the value being neither 0 nor negative;

for each list in the multiple lists, computing a score for each item within the list, wherein the score is a function of the associated rank within the list, and wherein the function is equal to 1/(n^(1/p));

for each item, summing corresponding scores from the multiple lists for a same item to produce a total score for the item;

ranking each item based on the corresponding total score of each item; and

outputting the combined list based on the ranking, wherein said first and second program instructions are stored on said computer readable media.

10. The computer program product of claim 9, further comprising:

third program instructions to assign a new value to the variable p to adjust the ranking in the combined list of items; and

fourth program instructions to repeat the combining,

wherein said third and fourth program instructions are stored on said computer readable media.

11. The computer program product of claim 9, wherein assigning the value to the variable p comprises automatically assigning the value to the variable p based on a predetermined business objective.

12. A method for providing a combined ranked list from multiple ranked lists, the method comprising:

receiving an identification of a domain of interest;

receiving information describing a business objective;

gathering multiple ranked lists on the domain of interest from multiple sources, wherein each list comprises items each having an associated rank, n, within the list;

combining the items in the multiple lists into a combined list, wherein the combining comprises;

automatically computing a score for each item within each of the multiple lists using a function equal to 1/(n^(1/p)), wherein p is a variable and the value of p is determined based on the business objective and is neither 0 nor negative;

for each item, combining the corresponding scores from the multiple lists for a same item to create a total score; and

combining the items into a combined list, wherein the rankings of the items with the combined list is based on the total score of each item; and

transmitting the combined list.

13. The method of claim 12 wherein the multiple sources have differing modalities.

14. The method of claim 13 wherein gathering the multiple ranked lists comprises:

gathering unstructured data indicating preferences for items within the domain of interest; and

constructing a ranked list from the unstructured data.

15. The method of claim 12 where the value of p is higher when the business objective indicates interest in ranking items with broad appeal higher than items with narrow appeal than when the business objective indicates interest in ranking items with narrow appeal higher than items with broad appeal.

16. The method of claim 12, further comprising:

calculating the associated rank n for each item within each list based on a position of the item with the corresponding list in the multiple lists.

17. The method of claim 12, further comprising:

receiving further information describing the business objective;

tuning the value of the variable p based on the further information;

repeating the combining based on the tuning; and

repeating the transmitting based on the repeating of the combining.