Systems and methods for identifying a named entity

• US 9,009,153 B2
• Filed: 03/31/2004
• Issued: 04/14/2015
• Est. Priority Date: 03/31/2004
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Khan, Omar Habib, Wang, Niniane, Lawrence, Stephen R.
• Primary Examiner(s): Mofiz, Apu
• Assistant Examiner(s): DAYE, CHELCIE L
• Application Number:
  US10/814,365
• Publication Number:
  US 20080005090A1
• Time in Patent Office: 4,031 Days
• Field of Search: 707/3, 707/5, 707/734, 707/750, 709/224, 725/53
• US Class Current: 707/734
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

Systems and methods for identifying a named entity are described. In one described system, a computer program, such as a query system, identifies an event associated with an article, identifies a named entity within the event, and creates an implicit search query comprising the named entity.

Claims

1. A computer-implemented method comprising:

determining a collection of named entity terms within a data store coupled to a computer of a user, a named entity term having an associated weight indicating a likelihood that the named entity term is included is a named entity;

identifying an event comprising a user interaction with an article on the computer;

identifying a plurality of named entity terms associated with the event;

identifying, from the collection the weight associated with each of the plurality of named entity terms;

responsive to identifying the weight associated with each of the plurality of named entity terms, automatically creating an implicit search query based at least in part on the plurality of named entity terms and the associated weights, the implicit search query based on a first named entity term with an associated weight indicating a higher likelihood that the first named entity term is a named entity more than on a second named entity term with an associated weight indicating a lower likelihood that the second named entity term is a named entity;

generating, from the computer, a plurality of search results relevant to the implicit search query; and

providing for display the generated plurality of search results.

2. The method of claim 1, wherein identifying a plurality of named entity terms associated with the event comprises identifying a term in the event that matches an entity term in the collection of named entity terms.

3. The method of claim 1, wherein determining the collection of named entity terms comprises monitoring instant messaging traffic.

4. The method of claim 1, wherein determining the collection of named entity terms comprises analyzing an email data store.

5. The method of claim 1, wherein determining the collection of named entity terms comprises analyzing a directory structure.

6. The method of claim 1, wherein determining the collection of named entity terms comprises searching a contact list.

7. The method of claim 1, wherein determining the collection of named entity terms comprises searching a news list.

8. The method of claim 1, wherein determining the collection of named entity terms comprises part of speech tagging.

9. The method of claim 1, wherein the named entity comprises one of an email address, an instant messaging name, and a proper noun.

10. The method of claim 1, further comprising storing the named entity in a user profile.

11. The method of claim 1, further comprising identifying a plurality of named entities for a name by using first name only, last name only, and combinations thereof.

12. The method of claim 11, further comprising filtering out at least one of the plurality of named entities having a high document frequency (DF).

13. The method of claim 1, further comprising:

receiving a result set associated with the implicit search query; and

providing the result set for display.

14. The method of claim 13, further comprising:

receiving an interest signal associated with one of the plurality of named entities; and

ranking the result set based at least in part on the interest signal.

30. The method of claim 1, wherein a frequency of each named entity term from the collection comprises an inverse document frequency of that named entity within the data store.

31. The method of claim 1, wherein a frequency of each named entity term comprises a term frequency of that named entity term within the data store.

32. The method of claim 1, wherein there is a plurality of types of named entities and further comprising:

assigning a weight to the implicit query based on a type of named entity on which the implicit query is based.

33. The method of claim 1, wherein determining the collection of named entity terms comprises:

storing names of people with whom the user communicates in the collection of named entity terms.

34. The method of claim 1, wherein determining the collection of named entity terms comprises:

filtering a collection of identified terms to remove common terms with a high document frequency; and

storing the filtered identified terms in the collection of named entity terms.

35. The method of claim 1, wherein determining the collection of named entity terms comprises:

adding nouns which appear less often in a set of documents stored in a repository to the collection of named entity terms.

36. The method of claim 1, further comprising:

storing a subset of the named entity terms in a first memory facilitating rapid lookup of the named entity terms.

37. The method of claim 1, wherein there is a plurality of types of named entities and wherein automatically creating the implicit search query comprises:

creating an implicit query for a specific type of data, the specific type of data determined responsive to a type of named entity on which the implicit query is based.

38. The method of claim 1, wherein the named entity term is included in a name of a person that is represented by the named entity.

39. The method of claim 1, wherein the named entity term is included in an email address that is represented by the named entity.

15. A computer-implemented method comprising:

determining a collection of named entity terms within a data store coupled to a computer of a user, a named entity term having an associated weight indicating a likelihood that the named entity term is included in a named entity;

receiving an event comprising a user interaction with an article on the computer;

identifying a plurality of named entity terms associated with the event;

identifying, from the collection, the weight associated with each of the plurality of named entity terms;

responsive to identifying the weight associated with each of the plurality of named entity terms, automatically creating an implicit query based at least in part on a first named entity term with an associated weight indicating a higher likelihood that the first named entity term is a named entity more than on a second named entity term with an associated weight indicating a lower likelihood that the second named entity term is a named entity;

transmitting the implicit query to a search engine;

receiving a result set from the search engine, the result set comprising one or more article identifiers; and

responsive to an associated score exceeding a threshold, outputting the one or more article identifiers.

16. A non-transitory computer-readable storage medium on which is encoded executable program code, the program code comprising:

program code for determining a collection of named entity terms within a data store coupled to a computer of a user, a named entity term having an associated weight indicating a likelihood that the named entity term is included in a named entity;

program code for identifying an event comprising a user interaction with an article on the computer;

program code for identifying a plurality of named entity terms associated with the event;

program code for identifying, from the collection, the weight associated with each of the plurality of named entity terms;

program code for automatically creating an implicit search query responsive to identifying the weight associated with each of the plurality of named entity terms, the implicit query based at least in part on the plurality of named entity terms and the associated weights, the implicit search query based on a first named entity term with an associated weight indicating a higher likelihood that the first named entity term is a named entity more than on a second named entity term with an associated weight indicating a lower likelihood that the second named entity term is a named entity;

program code for generating, from the computer, a plurality of search results relevant to the implicit search query; and

program code for providing for display the retrieved plurality of search results.

17. The computer-readable medium of claim 16, wherein program code for identifying a plurality of named entity terms associated with the event comprises program code for identifying a term in the event that matches an entity term in the collection of named entity terms.

18. The computer-readable medium of claim 16, further comprising program code for storing the named entity in a user profile.

19. The computer-readable medium of claim 16, wherein program code for determining the collection of named entity terms comprises program code for monitoring instant messaging traffic.

20. The computer-readable medium of claim 16, wherein program code for determining the collection of named entity terms comprises program code for analyzing an email data store.

21. The computer-readable medium of claim 16, wherein program code for determining the collection of named entity terms comprises program code for analyzing a directory structure.

22. The computer-readable medium of claim 16, wherein program code for determining the collection of named entity terms comprises program code for searching a contact list.

23. The computer-readable medium of claim 16, wherein program code for determining the collection of named entity terms comprises program code for searching a news list.

24. The computer-readable medium of claim 16, wherein program code for determining the collection of named entity terms comprises program code for part of speech tagging.

25. The computer-readable medium of claim 16, further comprising:

program code for receiving a result set associated with the implicit search query; and

program code for providing the result set for display.

26. The computer-readable medium of claim 25, further comprising:

program code for receiving an interest signal associated with one of the plurality of named entities; and

program code for ranking the result set based at least in part on the interest signal.

27. The computer-readable medium of claim 26, further comprising program code for identifying a plurality of named entities for a name by using first name only, last name only, and combinations thereof.

28. The computer-readable medium of claim 26, further comprising program code for filtering out at least one of the plurality of named entities having a high document frequency (DF).

29. A non-transitory computer-readable storage medium on which is encoded executable program code, the program code comprising:

program code for determining a collection of named entity terms within a data store coupled to a computer of a user, a named entity term having an associated weight indicating a likelihood that the named entity term is included in a named entity;

program code for receiving an event comprising a user interaction with an article on the computer;

program code for identifying a plurality of named entity terms associated with the event;

program code for identifying, from the collection, the weight associated with each of the plurality of named entity terms;

program code for automatically creating an implicit query, responsive to identifying the weight associated with each of the plurality of named entity terms, the implicit query based at least in part on a first named entity term with an associated weight indicating a higher likelihood that the first named entity term is a named entity more than on a second named entity term with an associated weight indicating a lower likelihood that the second named entity term is a named entity;

program code for transmitting the implicit query to a search engine;

program code for receiving a result set from the search engine, the result set comprising one or more article identifiers; and

responsive to an associated score exceeding a threshold level of relevance, outputting the one or more article identifiers.