Ranking expertise

• US 8,892,549 B1
• Filed: 06/29/2007
• Issued: 11/18/2014
• Est. Priority Date: 06/29/2007
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Thakur, Shashidhar A.
• Primary Examiner(s): Ahn, Sangwoo
• Application Number:
  US11/772,022
• Time in Patent Office: 2,699 Days
• US Class Current: 707/728
• CPC Class Codes: G06F 16/3331 Query processing
• CPC Class Codes: G06F 16/338 Presentation of query results
• CPC Class Codes: G06Q 10/10 Office automation; Time management

Abstract

Methods, systems and apparatus, including computer program products, for ranking expertise. In some implementations a method is provided that includes identifying a plurality of identities, and identifying a plurality of topics using one or more documents in a repository. For a document in a corpus of documents identifying one or more occurrences of any identity in the plurality of identities and one or more occurrences of any topic in the plurality of topics, determining an association between the identities occurring in the document and the document including deriving an identity score for each unique identity occurring in the document, determining an association between the topics occurring in the document and the document including deriving a topic score for each unique topic occurring in the document, and using the determined associations to derive a score of the document with respect to identities and topics occurring in the document.

Claims

1. A computer-implemented method comprising:

identifying a plurality of identities stored in a repository of identities, each identity corresponding to an expert of one or more topics;

identifying a plurality of topics stored in a repository of information, each topic i) describing information included by a document of a corpus of documents and ii) distinguishing the document from the remaining documents of the corpus of documents, the plurality of topics including the one or more topics; and

processing each document in the corpus of documents, the processing including;

identifying one or more identities of the plurality of identities that occur within the document,

identify one or more topics of the plurality of topics that occur within the document,

for each identity that occurs within the document, determining, using one or more processors, an identity score for the identity with respect to the document, the identity score indicating a degree of relevance between the associated identity and the document,

for each topic that occurs within the document, determining a topic score for the topic with respect to the document, the topic score indicating a degree of relevance between the associated topic and the document,

identifying one or more combinations of i) the one or more identities that occur within the document, and ii) the one or more topics that occur within the document,

for each identified combination, determining an aggregate score for the document based on the identity score associated with the combination for the document and the topic score associated with the combination for the document; and

aggregating, for each identified combination, the aggregate score of each document of the corpus of documents for the identified combination to define a composite score of the identified combination across the corpus of documents.

2. The method of claim 1, where identifying a plurality of topics comprises:

identifying a plurality of n-grams from each document in the corpus of documents;

determining an inverse document frequency (IDF) score for each n-gram; and

sorting the n-grams by IDF score to identify topics corresponding to one or more of the n-grams.

3. The method of claim 1, where determining the identity score for the identity with respect to the document includes determining the identity score based on features of a reference to the identity occurring in the document.

4. The method of claim 3, where features of a reference to the identity include one or more of:

a location within the document of the reference; and

typographical properties of the reference.

5. The method of claim 1, where determining the identity score for the identity with respect to the document includes determining the identity score based on a frequency of references to the identity occurring in the document.

6. The method of claim 1, where determining the topic score for the topic with respect to the document includes determining the topic score based on features of occurrence of the topic in the document.

7. The method of claim 6, where features of an occurrence include one or more of:

a location within the document of the occurrence; and

typographical properties of the occurrence.

40. The method of claim 1, wherein for each identified combination, multiplying the identity score associated with the combination for the document and the topic score associated with the combination for the document.

8. A method comprising:

receiving a search query;

determining, using one or more processors, that the search query matches an identity among a plurality of identities, each identity of the plurality of identities corresponding to an expert of one or more experts;

identifying multiple topics associated with the identity based on multiple documents in which the identity is referred, each topic of the multiple topics i) describing information included by a document of the multiple documents and ii) distinguishing the document from the remaining documents of the multiple documents;

for each document of the multiple documents, obtaining a composite score of each topic of the document based on an aggregation of aggregate scores of the topic for the document, the aggregate score of each topic based on an identity score associated with the document and a topic score of each topic associated with the document, the topic score for each topic indicating a degree of relevance between the topic and the document, and the identity score for the identity indicating a degree of relevance between the identity and the document;

ordering the multiple topics based on a respective composite score associated with each document of the multiple documents; and

presenting the multiple topics based on the ordering.

9. The method of claim 8, where presenting the multiple topics includes:

presenting, for each of the multiple topics, one or more of the multiple documents, the one or more of the multiple documents being documents most relevant to the corresponding topic.

10. The method of claim 8, further comprising:

identifying the plurality of identities from one or more directories.

41. The method of claim 8, wherein the aggregate score of each topic for each document is a product of the topic score of the topic associated with the document and the identity score associated with the document.

11. A method comprising:

receiving a search query;

determining, using one or more processors, that the search query matches a topic among a plurality of topics, each topic of the plurality of topics i) describing information included by a document of multiple documents and ii) distinguishing the document from the remaining documents of the multiple documents;

identifying multiple identities associated with the topic based on multiple documents in which the topic is referred, each identity of the multiple identities corresponding to an expert of one or more experts;

for each document of the multiple documents, obtaining a composite score of each identity of the document based on an aggregation of aggregate scores of the identity for the document, the aggregate score of each identity based on a topic score associated with the document and an identity score of each identity associated with the document, the identity score for each identity indicating a degree of relevance between the identity and the document, and the topic score for the topic indicating a degree of relevance between the topic and the document;

ordering the multiple identities based on a respective composite score associated with each document of the multiple documents; and

presenting the multiple identities based on the ordering.

12. The method of claim 11, where presenting the multiple identities includes:

presenting, for each of the multiple identities, one or more of the multiple documents, the one or more of the multiple documents being documents most relevant to the corresponding identity.

13. The method of claim 11, further comprising:

identifying the plurality of topics from a corpus of documents.

42. The method of claim 11, wherein aggregate score of each topic for each document is a product of the topic score of the topic associated with the document and the identity score associated with the document.

14. A computer program product, encoded on a non-transitory computer-readable medium, operable to cause data processing apparatus to perform operations comprising:

identifying a plurality of identities stored in a repository of identities, each identity corresponding to an expert of one or more topics;

identifying a plurality of topics stored in a repository of information, each topic i) describing information included by a document of a corpus of documents and ii) distinguishing the document from the remaining documents of the corpus of documents, the plurality of topics including the one or more topics; and

processing each document in the corpus of documents, the processing including;

identifying one or more identities of the plurality of identities that occur within the document,

identifying one or more topics of the plurality of topics that occur within the document,

for each identity that occurs within the document, determining an identity score for the identity with respect to the document, the identity score indicating a degree of relevance between the associated identity and the document,

for each topic that occurs within the document, determining a topic score for the topic with respect to the document, the topic score indicating a degree of relevance between the associated topic and the document,

identifying one or more combinations of i) the one or more identities that occur within the document, and ii) the one or more topics that occur within the document,

for each identified combination, determining an aggregate score for the document based on the identity score associated with the combination for the document and the topic score associated with the combination for the document; and

aggregating, for each identified combination, the aggregate score of each document of the corpus of documents for the identified combination to define a composite score of the identified combination across the corpus of documents.

15. The program product of claim 14, where identifying a plurality of topics comprises:

identifying a plurality of n-grams from each document in the corpus of documents;

determining an inverse document frequency (IDF) score for each n-gram; and

sorting the n-grams by IDF score to identify topics corresponding to one or more of the n-grams.

16. The program product of claim 14, where determining the identity score for the identity with respect to the document includes determining the identity score based on features of a reference to the identity occurring in the document.

17. The program product of claim 16, where features of a reference to the identity include one or more of:

a location within the document of the reference; and

typographical properties of the reference.

18. The program product of claim 14, where determining the identity score for the identity with respect to the document includes determining the identity score based on a frequency of references to the identity occurring in the document.

19. The program product of claim 14, where determining the topic score for the topic with respect to the document includes determining the topic score based on features of occurrence of the topic in the document.

20. The program product of claim 19, where features of an occurrence include one or more of:

a location within the document of the occurrence; and

typographical properties of the occurrence.

21. A computer program product, encoded on a non-transitory computer-readable medium, operable to cause data processing apparatus to perform operations comprising:

receiving a search query;

determining that the search query matches an identity among a plurality of identities, each identity of the plurality of identities corresponding to an expert of one or more experts;

identifying multiple topics associated with the identity based on multiple documents in which the identity is referred, each topic of the multiple topics i) describing information included by a document of the multiple documents and ii) distinguishing the document from the remaining documents of the multiple documents;

for each document of the multiple documents, obtaining a composite score of each topic of the document based on an aggregation of aggregate scores of the topic for the document, the aggregate score of each topic based on an identity score associated with the document and a topic score of each topic associated with the document, the topic score for each topic indicating a degree of relevance between the topic and the document, and the identity score for the identity indicating a degree of relevance between the identity and the document;

ordering the multiple topics based on a respective composite score associated with each document of the multiple documents; and

presenting the multiple topics based on the ordering.

22. The program product of claim 21, where presenting the multiple topics includes:

presenting, for each of the multiple topics, one or more of the multiple documents, the one or more of the multiple documents being documents most relevant to the corresponding topic.

23. The program product of claim 21, further operable to cause data processing apparatus to perform operations comprising:

identifying the plurality of identities from one or more directories.

24. A computer program product, encoded on a non-transitory computer-readable medium, operable to cause data processing apparatus to perform operations comprising:

receiving a search query;

determining that the search query matches a topic among a plurality of topics, each topic of the plurality of topics i) describing information included by a document of multiple documents and ii) distinguishing the document from the remaining documents of the multiple documents;

identifying multiple identities associated with the topic based on multiple documents in which the topic is referred, each identity of the multiple identities corresponding to an expert of one or more experts;

for each document of the multiple documents, obtaining a composite score of each identity of the document based on an aggregation of aggregate scores of the identity for the document, the aggregate score of each identity based on a topic score associated with the document and an identity score of each identity associated with the document, the identity score for each identity indicating a degree of relevance between the identity and the document, and the topic score for the topic indicating a degree of relevance between the topic and the document;

ordering the multiple identities based on a respective composite score associated with each document of the multiple documents; and

presenting the multiple identities based on the ordering.

25. The program product of claim 24, where presenting the multiple identities includes:

presenting, for each of the multiple identities, one or more of the multiple documents, the one or more of the multiple documents being documents most relevant to the corresponding identity.

26. The program product of claim 24, further operable to cause data processing apparatus to perform operations comprising:

identifying the plurality of topics from a corpus of documents.

28. The system of claim 26, where identifying a plurality of topics comprises:

identifying a plurality of n-grams from each document in the corpus of documents;

determining an inverse document frequency (IDF) score for each n-gram; and

sorting the n-grams by IDF score to identify topics corresponding to one or more of the n-grams.

29. The system of claim 26, where determining the identity score for the identity with respect to the document includes determining the identity score based on features of a reference to the identity occurring in the document.

30. The system of claim 29, where features of a reference to the identity include one or more of:

a location within the document of the reference; and

typographical properties of the reference.

31. The system of claim 26, where determining the identity score for the identity with respect to the document includes determining the identity score based on a frequency of references to the identity occurring in the document.

32. The system of claim 26, where determining the topic score for the topic with respect to the document includes determining the topic score based on features of occurrence of the topic in the document.

33. The system of claim 32, where features of an occurrence include one or more of:

a location within the document of the occurrence; and

typographical properties of the occurrence.

27. A system comprising:

one or more processors configured to perform operations including;

identifying a plurality of identities stored in a repository of identities, each identity corresponding to an expert of one or more topics;

identifying a plurality of topics stored in a repository of information, each topic i) describing information included b a document of a corpus of documents and ii) distinguishing the document from the remaining documents of the corpus of documents, the plurality of topics including the one or more topics; and

processing each document in the corpus of documents, the processing including;

means for identifying one or more identities of the plurality of identities that occur within the document,

means for identifying one or more topics of the plurality of topics that occur within the document,

for each identity occurring within the document, means for determining an identity score for the identity with respect to the document, the identity score indicating a degree of relevance between the associated identity and the document,

for each topic occurring within the document, means for determining a topic score for the topic with respect to the document, the topic score indicating a degree of relevance between the associated topic and the document,

means for identifying one or more combinations of i) the one or more identities that occur within the document, and ii) the one or more topics that occur within the document,

for each identified combination, means for determining an aggregate score for the document based on the identity score associated with the combination for the document and the topic score associated with the combination for the document; and

means for aggregating, for each identified combination, the aggregate score of each document of the corpus of documents for the identified combination to define a composite score of the identified combination across the corpus of documents.

34. A system comprising:

one or more processors configured to perform operations including;

receiving a search query;

determining that the search query matches an identity among a plurality of identities, each identity of the plurality of identities corresponding to an expert of one or more experts;

identifying multiple topics associated with the identity based on multiple documents in which the identity is referred, each topic of the multiple topics i) describing information included by a document of the multiple documents and ii) distinguishing the document from the remaining documents of the multiple documents;

for each document of the multiple documents, obtaining a composite score of each topic of the document based on an aggregation of aggregate scores of the topic for the document, the aggregate score of each topic based on an identity score associated with the document and a topic score of each topic associated with the document, the topic score for each topic indicating a degree of relevance between the topic and the document, and the identity score for the identity indicating a degree of relevance between the identity and the document;

ordering the multiple topics based on a respective composite score associated with each document of the multiple documents; and

presenting the multiple topics based on the ordering.

35. The system of claim 34, where presenting the multiple topics includes:

presenting, for each of the multiple topics, one or more of the multiple documents, the one or more of the multiple documents being documents most relevant to the corresponding topic.

36. The system of claim 34, wherein the operations further comprise:

identifying the plurality of identities from one or more directories.

37. A system comprising:

one or more processors configured to perform operations including;

receiving a search query;

determining that the search query matches a topic among a plurality of topics, each topic of the plurality of topics i) describing information included by a document of multiple documents and ii) distinguishing the document from the remaining documents of the multiple documents;

identifying multiple identities associated with the topic based on multiple documents in which the topic is referred, each identity of the multiple identities corresponding to an expert of one or more experts;

for each document of the multiple documents, obtaining a composite score of each identity of the document based on an aggregation of aggregate scores of the identity for the document, the aggregate score of each identity based on a topic score associated with the document and an identity score of each identity associated with the document, the identity score for each identity indicating a degree of relevance between the identity and the document, and the topic score for the topic indicating a degree of relevance between the topic and the document;

ordering the multiple identities based on a respective composite score associated with each document of the multiple documents; and

presenting the multiple identities based on the ordering.

38. The system of claim 37, where presenting the multiple identities includes:

presenting, for each of the multiple identities, one or more of the multiple documents, the one or more of the multiple documents being documents most relevant to the corresponding identity.

39. The system of claim 37, wherein the operations further comprise:

identifying the plurality of topics from a corpus of documents.