Techniques for ranking search results

• US 8,903,810 B2
• Filed: 10/16/2008
• Issued: 12/02/2014
• Est. Priority Date: 12/05/2005
• Status: Expired due to Fees

• Current Assignee: Collarity Incorporated
• Original Assignee: Collarity Incorporated
• Inventors: Ismalon, Emil
• Primary Examiner(s): Mahmoudi, Tony
• Assistant Examiner(s): PHAN, TUANKHANH D
• Application Number:
  US12/253,087
• Publication Number:
  US 20090119261A1
• Time in Patent Office: 2,238 Days
• Field of Search: 707/732, 707/748, 707/705, 382/159
• US Class Current: 707/723
• CPC Class Codes: G06F 16/3322 using system suggestions G06F16/3325 takes precedence
• CPC Class Codes: G06F 16/355 Class or cluster creation or modification
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation

Abstract

A computer-implemented method is provided, which includes presenting to a user (30) a range of levels of personalization of search results, including a personalized level, a global level that is not personalized, and a community level between the personalized level and the global level. An indication (550) of a desired one of the levels, and a search query (52) consisting of one or more query terms, are received from the user (30). Responsively to the search query, a search result listing (54) is generated. At least a portion of the search result listing (54) is ranked at least in part responsively to the indication (550), and at least a portion of the ranked search result (54) listing is presented to the user (30). Other embodiments are also described.

Claims

1. A computer-implemented method comprising:

presenting, by a search system, to a user who has a personal profile in the search system and belongs in the search system to one or more communities, a range of levels of personalization of search results, including a personalized level, a global level that is not personalized, and a community level between the personalized level and the global level;

receiving from the user, by the search system, an indication of a desired preference that a mixture of two or more of the levels be used for ranking search results;

regardless of whether the mixture includes the community level of personalization, maintaining in the search system that the user belongs to the one or more communities;

receiving from the user, by the search system, a search query comprising one or more query terms;

responsively to the search query, generating, by the search system, a search result listing;

ranking, by the search system, at least a portion of the search result listing at least in part responsively to the indication, including, if the mixture includes the community level of personalization, ranking the at least a portion of the search result listing in part based on characteristics of the one or more communities to which the user belongs; and

presenting, by the search system, at least a portion of the ranked search result listing to the user.

2. The method according to claim 1, wherein ranking comprises, if the received indication indicates the personalized level of personalization:

calculating respective correlation scores between the personal profile and a plurality of search result documents of the at least a portion of the search result listing; and

ranking the at least a portion of the search result listing at least in part responsively to the correlation scores.

3. The method according to claim 2, wherein the personal profile includes a personal association graph (PAG), and wherein the method further comprises constructing the PAG including as vertices at least a portion of previous query terms of previous searches conducted by the user.

4. The method according to claim 1, wherein the user is one of a plurality of users, and wherein ranking comprises, if the received indication indicates the community level of personalization:

constructing a topic profile that represents interactions of a plurality of previous searches conducted by the plurality of users within a single topic;

calculating respective correlation scores between the topic profile and a plurality of search result documents of the at least a portion of the search result listing; and

ranking the at least a portion of the search result listing at least in part responsively to the correlation scores.

5. The method according to claim 4, wherein the topic profile includes a topic association graph (TAG), wherein each of the previous searches includes one or more previous query terms, and wherein constructing the topic profile comprises constructing the TAG including at least a portion of the previous query terms as vertices.

6. The method according to claim 1, wherein the user is one of a plurality of users, and wherein ranking comprises, if the received indication indicates the community level of personalization:

constructing a plurality of personal profiles for respective users in the plurality of users, each of which personal profiles represents interactions of a plurality of previous searches conducted by one of the users with a respective plurality of previous search result documents presented to the one of the users during the previous searches;

identifying a subset of the personal profiles as correlated with one another;

constructing a group profile by combining at least a portion of each of the personal profiles in the subset;

calculating respective correlation scores between the group profile a plurality of search result documents of the at least a portion of the search result listing; and

ranking the at least a portion of the search result listing at least in part responsively to the correlation scores.

7. The method according to claim 6,

wherein the personal profiles include respective personal association graph (PAGs),

wherein each of the previous searches includes one or more first query terms,

wherein constructing the personal profiles comprises constructing the PAGs including at least a portion of the first query terms as vertices,

wherein the group profile includes a group association graph (GRAG), and

wherein constructing the group profile comprises constructing the GRAG by combining at least a portion of each of the PAGs in the subset.

8. The method according to claim 7, wherein generating the search result listing comprises generating the search result listing responsively to the query terms and at least one of the previous query terms in the GRAG linked to all of the query terms in the GRAG.

9. The method according to claim 1, wherein the user is one of a plurality of users, and wherein ranking comprises:

constructing a plurality of document profiles associated with a respective plurality of search result documents, each of which document profiles represents interactions of (a) a plurality of previous searches conducted by the plurality of users with (b) the respective search result document associated with the document profile; and

ranking the at least a portion of the search result listing at least in part responsively to the indication and respective correlation scores between the query terms and at least a portion of the document profiles.

10. The method according to claim 9, wherein the plurality of document profiles includes a respective plurality of document association graphs (DAGs), wherein the previous searches include one or more respective previous query terms, and wherein constructing the plurality of document profiles comprises constructing the plurality of DAGs, each of which includes one or more of the previous query terms as vertices.

11. The method according to claim 1, wherein receiving the indication of the desired preference that the mixture of the two or more of the levels be used for ranking the search results comprises receiving an indication of positioning, by the user, of a sliding pointer between two of the levels.

12. A computer-implemented method comprising:

constructing, by a search system, a topic association graph (TAG) that includes terms as vertices, each of which vertices is linked to one or more other vertices by respective edges having respective edge scores;

presenting, by the search system, to a user who has a personal profile in the search system and belongs in the search system to one or more communities, a range of levels of personalization of search results, including a personalized level, a global level that is not personalized, and a community level between the personalized level and the global level;

receiving from the user, by the search system, an indication of a desired preference that a mixture of two or more of the levels be used for ranking search results;

regardless of whether the mixture includes the community level of personalization, maintaining in the search system that the user belongs to the one or more communities;

receiving, by the search system, from the user a search query, which comprises one or more terms included in a portion of the vertices of the TAG;

responsively to the search query, generating, by the search system, a search result listing including a plurality of documents; and

if the mixture includes the community level of personalization;

generating, by the search system, one or more subgraphs of the TAG, each of which subgraphs consists of all of the terms of the search query plus one or more terms of the TAG that are most highly linked in the TAG to all of the terms of the search query;

ranking, by the search system, at least a portion of the search result listing at least in part responsively to a comparison between (a) the one or more subgraphs of the TAG and (b) at least a portion of the documents of the search result listing; and

presenting, by the search system, at least a portion of the ranked search result listing to the user.

13. The method according to claim 12, wherein the user is one of a plurality of users, and wherein constructing the TAG comprises:

constructing respective personal profiles for each of a portion of the users, each of which profiles represents interactions of a plurality of previous searches conducted by the user with a respective plurality of previous search result documents presented to the user during the previous searches; and

combining respective portions of the personal profiles, which portions are associated with the topic, to form the TAG.

14. The method according to claim 12, wherein generating the one or more subgraphs of the TAG comprises:

calculating respective association scores of candidate subgraphs, which candidate subgraphs consist of the terms of the search query plus respective candidate terms directly linked to all of the term of the search query in the TAG, which association scores represent strengths of association among the vertices of the respective candidate subgraphs; and

selecting the candidate subgraphs having the highest association scores as the one or more subgraphs of the TAG.

15. Apparatus comprising a search system, which comprises:

an interface for communicating with a user; and

a processor, which is configured to present to the user who has a personal profile in the search system and belongs in the search system to one or more communities, via the interface, a range of levels of personalization of search results, including a personalized level, a global level that is not personalized, and a community level between the personalized level and the global level;

receive from the user, via the interface, an indication of a desired preference that a mixture of two or more of the levels be used for ranking search results;

regardless of whether the mixture includes the community level of personalization, maintain in the search system that the user belongs to the one or more communities;

receive from the user, via the interface, a search query consisting of one or more query terms;

responsively to the search query, generate a search result listing;

rank at least a portion of the search result listing at least in part responsively to the indication, including, if the mixture includes the community level of personalization, ranking the at least a portion of the search result listing in part based on characteristics of the one or more communities to which the user belongs; and

present at least a portion of the ranked search result listing to the user via the interface.

16. The apparatus according to claim 15, wherein the processor is configured to, if the received indication indicates the personalized level of personalization, calculate respective correlation scores between the personal profile and a plurality of search result documents of the at least a portion of the search result listing, and rank the at least a portion of the search result listing at least in part responsively to the correlation scores.

17. The apparatus according to claim 15, wherein the user is one of a plurality of users, and wherein the processor is configured to, if the received indication indicates the community level of personalization, construct a topic profile that represents interactions of a plurality of previous searches conducted by the plurality of users within a single topic, calculate respective correlation scores between the topic profile and a plurality of search result documents of the at least a portion of the search result listing, and rank the at least a portion of the search result listing at least in part responsively to the correlation scores.

18. The apparatus according to claim 15, wherein the user is one of a plurality of users, and wherein the processor is configured to, if the received indication indicates the community level of personalization, construct a plurality of personal profiles for respective users in the plurality of users, each of which personal profiles represents interactions of a plurality of previous searches conducted by one of the users with a respective plurality of previous search result documents presented to the one of the users during the previous searches;

identify a subset of the personal profiles as correlated with one another;

construct a group profile by combining at least a portion of each of the personal profiles in the subset;

calculate respective correlation scores between the group profile a plurality of search result documents of the at least a portion of the search result listing; and

rank the at least a portion of the search result listing at least in part responsively to the correlation scores.

19. The apparatus according to claim 15, wherein the user is one of a plurality of users, wherein the interface is configured to communicate with the plurality of users, and wherein the processor is configured to construct a plurality of document profiles associated with a respective plurality of search result documents, each of which document profiles represents interactions of (a) a plurality of previous searches conducted by the plurality of users with (b) the respective search result document associated with the document profile, and rank the at least a portion of the search result listing at least in part responsively to the indication and respective correlation scores between the query terms and at least a portion of the document profiles.

20. The apparatus according to claim 15, wherein the processor is configured to receive the indication of the desired preference that the mixture of the two or more of the levels be used for ranking the search results by receiving an indication of positioning, by the user, of a sliding pointer between two of the levels.

21. Apparatus comprising a search system, which comprises:

an interface for communicating with a user; and

a processor, which is configured to construct a topic association graph (TAG) that includes terms as vertices, each of which vertices is linked to one or more other vertices by respective edges having respective edge scores;

present to a user who has a personal profile in the search system and belongs in the search system to one or more communities, a range of levels of personalization of search results, including a personalized level, a global level that is not personalized, and a community level between the personalized level and the global level;

receive from the user an indication of a desired preference that a mixture of two or more of the levels be used for ranking search results;

regardless of whether the mixture includes the community level of personalization, maintain in the search system that the user belongs to the one or more communities;

receive from the user, via the interface, a search query, which comprises one or more terms included in a portion of the vertices of the TAG;

responsively to the search query, generate a search result listing including a plurality of documents; and

, if the mixture includes the community level of personalization;

(i) generate one or more subgraphs of the TAG, each of which subgraphs consists of all of the terms of the search query plus one or more terms of the TAG that are most highly linked in the TAG to all of the terms of the search query;

(ii) rank at least a portion of the search result listing at least in part responsively to a comparison between (a) the one or more subgraphs of the TAG and (b) at least a portion of the documents of the search result listing; and

(iii) present at least a portion of the ranked search result listing to the user via the interface.

22. The apparatus according to claim 21, wherein the user is one of a plurality of users, and wherein the processor is configured to construct the TAG by:

constructing respective personal profiles for each of a portion of the users, each of which profiles represents interactions of a plurality of previous searches conducted by the user with a respective plurality of previous search result documents presented to the user during the previous searches; and

combining respective portions of the personal profiles, which portions are associated with the topic, to form the TAG.

23. The apparatus according to claim 21, wherein the processor is configured to generate the one or more subgraphs of the TAG by:

calculating respective association scores of candidate subgraphs, which candidate subgraphs consist of the terms of the search query plus respective candidate terms directly linked to all of the term of the search query in the TAG, which association scores represent strengths of association among the vertices of the respective candidate subgraphs; and

selecting the candidate subgraphs having the highest association scores as the one or more subgraphs of the TAG.

24. A computer software product, the product comprising a tangible computer-readable non-transitory medium in which program instructions are stored, which instructions, when read by a computer of a search system, cause the computer to present to a user who has a personal profile in the search system and belongs in the search system to one or more communities a range of levels of personalization of search results, including a personalized level, a global level that is not personalized, and a community level between the personalized level and the global level;

receive from the user an indication of a desired preference that a mixture of two or more of the levels be used for ranking search results;

regardless of whether the mixture includes the community level of personalization, maintain in the search system that the user belongs to the one or more communities;

receive from the user a search query comprising one or more query terms;

responsively to the search query, generate a search result listing;

rank at least a portion of the search result listing at least in part responsively to the indication, including, if the mixture includes the community level of personalization, ranking the at least a portion of the search result listing in part based on characteristics of the one or more communities to which the user belongs; and

present at least a portion of the ranked search result listing to the user.

25. The computer software product according to claim 24, wherein the instructions, when read by the computer, cause the computer to receive the indication of the desired preference that the mixture of the two or more of the levels be used for ranking the search results by receiving an indication of positioning, by the user, of a sliding pointer between two of the levels.

26. A computer software product, the product comprising a tangible computer-readable non-transitory medium in which program instructions are stored, which instructions, when read by a computer of a search system, cause the computer to construct a topic association graph (TAG) that includes terms as vertices, each of which vertices is linked to one or more other vertices by respective edges having respective edge scores;

present to a user who has a personal profile in the search system and belongs in the search system to one or more communities, a range of levels of personalization of search results, including a personalized level, a global level that is not personalized, and a community level between the personalized level and the global level;

receive from the user an indication of a desired preference that a mixture of two or more of the levels be used for ranking search results;

regardless of whether the mixture includes the community level of personalization, maintain in the search system that the user belongs to the one or more communities;

receive from the user a search query, which comprises one or more terms included in a portion of the vertices of the TAG;

responsively to the search query, generate a search result listing including a plurality of documents; and

, if the mixture includes the community level of personalization;

(i) generate one or more subgraphs of the TAG, each of which subgraphs consists of all of the terms of the search query plus one or more terms of the TAG that are most highly linked in the TAG to all of the terms of the search query;

(ii) rank at least a portion of the search result listing at least in part responsively to a comparison between (a) the one or more subgraphs of the TAG and (b) at least a portion of the documents of the search result listing; and

(iii) present at least a portion of the ranked search result listing to the second user.

27. The computer software product according to claim 26, wherein the instructions, when read by the computer, cause the computer to generate the one or more subgraphs of the TAG by:

calculating respective association scores of candidate subgraphs, which candidate subgraphs consist of the terms of the search query plus respective candidate terms directly linked to all of the term of the search query in the TAG, which association scores represent strengths of association among the vertices of the respective candidate subgraphs; and

selecting the candidate subgraphs having the highest association scores as the one or more subgraphs of the TAG.

28. A computer-implemented method comprising:

presenting, by a search system, to a user who has a personal profile in the search system and belongs in the search system to one or more communities, a range of levels of personalization of search results, including a personalized level, a global level that is not personalized, and a community level between the personalized level and the global level;

receiving from the user, by the search system, an indication of a desired preference that two or more of the levels be used for ranking the search results;

regardless of whether the desired preference includes the community level of personalization, maintaining in the search system that the user belongs to the one or more communities;

receiving from the user, by the search system, a search query comprising one or more query terms;

responsively to the search query, generating, by the search system, a search result listing;

ranking, by the search system, at least a portion of the search result listing at least in part responsively to a mixture of the two or more user-indicated levels, including, if the mixture includes the community level of personalization, ranking the at least a portion of the search result listing in part based on characteristics of the one or more communities to which the user belongs; and

presenting, by the search system, at least a portion of the ranked search result listing to the user.

29. Apparatus comprising a search system, which comprises:

an interface for communicating with a user; and

a processor, which is configured to present to the user who has a personal profile in the search system and belongs in the search system to one or more communities, via the interface, a range of levels of personalization of search results, including a personalized level, a global level that is not personalized, and a community level between the personalized level and the global level;

receive from the user, via the interface, an indication of the desired preference by receiving an indication of a desired preference that two or more of the levels be used for ranking the search results;

regardless of whether the desired preference includes the community level of personalization, maintain in the search system that the user belongs to the one or more communities;

rank the at least a portion of the search result listing at least in part responsively to a mixture of the two or more user-indicated levels, including, if the mixture includes the community level of personalization, ranking the at least a portion of the search result listing in part based on characteristics of the one or more communities to which the user belongs; and

present at least a portion of the ranked search result listing to the user via the interface.

30. A computer software product, the product comprising a tangible computer-readable non-transitory medium in which program instructions are stored, which instructions, when read by the computer, cause the computer to present, to a user who has a personal profile in the search system and belongs in the search system to one or more communities, a range of levels of personalization of search results, including a personalized level, a global level that is not personalized, and a community level between the personalized level and the global level;

receive from the user an indication of a desired preference that two or more of the levels be used for ranking the search results;

regardless of whether the desired preference includes the community level of personalization, maintain in the computer that the user belongs to the one or more communities;

receive from the user a search query comprising one or more query terms;

responsively to the search query, generate a search result listing;

rank at least a portion of the search result listing at least in part responsively to a mixture of the two or more user-indicated levels, including, if the mixture includes the community level of personalization, ranking the at least a portion of the search result listing in part based on characteristics of the one or more communities to which the user belongs; and

present at least a portion of the ranked search result listing to the user.