Identification of semantic units from within a search query

• US 8,719,262 B1
• Filed: 09/14/2012
• Issued: 05/06/2014
• Est. Priority Date: 10/04/2000
• Status: Expired due to Fees

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google Inc. (Alphabet Inc.)
• Inventors: Bharat, Krishna, Ghemawat, Sanjay, Hoelzle, Urs
• Primary Examiner(s): Morrison, Jay
• Assistant Examiner(s): GORTAYO, DANGELINO N
• Application Number:
  US13/616,094
• Time in Patent Office: 599 Days
• Field of Search: 707/1, 707/3, 707/5, 707/7, 707/100, 707/706, 707/713, 707/723, 707/728, 707/759, 707/765, 707/771, 707/779
• US Class Current: 707/728
• CPC Class Codes: G06F 16/3334 Selection or weighting of terms from queries, including natural language queries
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: Y10S 707/99931 Database or file accessing
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

A search engine for searching a corpus improves the relevancy of the results by classifying multiple terms in a search query as a single semantic unit. A semantic unit locator of the search engine generates a subset of documents that are generally relevant to the query based on the individual terms within the query. Combinations of search terms that define potential semantic units from the query are then evaluated against the subset of documents to determine which combinations of search terms should be classified as a semantic unit. The resultant semantic units are used to refine the results of the search.

Claims

1. A method comprising:

identifying, by one or more devices, documents relating to a query;

generating, by the one or more devices, a plurality of substrings from the query;

calculating, by the one or more devices and for a particular substring of the plurality of substrings, a value that corresponds to a comparison between the identified documents and the particular substring;

determining, by the one or more devices, that the calculated value for the particular substring satisfies a particular threshold associated with identifying compounds;

selecting, by the one or more devices and from two or more of the plurality of substrings, the particular substring as a semantic unit based on the calculated value for the particular substring satisfying the particular threshold; and

obtaining, by the one or more devices, a refined list of documents by refining the identified documents based on the semantic unit.

2. The method of claim 1, where identifying the documents includes:

generating an initial list of relevant documents; and

selecting a predetermined number of documents, in the initial list of relevant documents, as the identified documents.

3. The method of claim 1, further including:

discarding, before selecting the particular string, one or more substrings of the plurality of substrings that overlap with one or more other substrings of the plurality of substrings.

4. The method of claim 1, where the value is weighted based on a ranking defined by relevance of the identified documents.

5. The method of claim 1, where identifying the documents includes:

identifying the documents relating to the query by comparing search terms in the query to an index of a corpus.

6. The method of claim 1, further comprising:

providing information based on the refined list of documents.

7. A non-transitory computer-readable medium storing instructions, the instructions comprising:

one or more instructions that, when executed by at least one processor, cause the at least one processor to;

receive a search query;

identify documents based on the search query;

generate a plurality of substrings based on the search query;

calculate, for a particular substring of the plurality of substrings, a value based on the identified documents and the particular substring;

determine that the calculated value for the particular substring satisfies a particular threshold associated with identifying compounds;

select the particular substring, as a semantic unit, from the plurality of substrings based on the calculated value for the particular substring satisfying the particular threshold; and

refine the identified documents based on the semantic unit.

8. The non-transitory computer-readable medium of claim 7, where the one or more instructions to calculate the value comprise:

one or more instructions that, when executed by the at least one processor, cause the at least one processor to;

identify, as a subset of documents, a predetermined number of most relevant documents of the identified documents, and

calculate the value based on one or more documents, of the subset of documents, that include the particular substring.

9. The non-transitory computer-readable medium of claim 7, where the instructions further comprise:

one or more instructions that, when executed by the at least one processor, cause the at least one processor to;

discard, before selecting the particular substring, one or more of the plurality of substrings that overlap with one or more other substrings of the plurality of substrings.

10. The non-transitory computer-readable medium of claim 7, where the value is weighted based on a ranking defined by relevance of the identified documents.

11. The non-transitory computer-readable medium of claim 7, where the one or more instructions to identify the documents comprise:

one or more instructions that, when executed by the at least one processor, cause the at least one processor to;

generate a list of relevant documents based on search terms of the search query, and

identify the documents, from the list of relevant documents, that are most relevant to the search query.

12. The non-transitory computer-readable medium of claim 7, where the value indicates a fraction of the identified documents that contain the particular substring.

13. The non-transitory computer-readable medium of claim 7, where the instructions further comprise:

one or more instructions that, when executed by the at least one processor, cause the at least one processor to;

provide information based on the identified documents after refining the identified documents based on the semantic unit.

14. A system comprising:

a server, including a processor, to;

identify documents relating to a query;

generate a plurality of substrings from the query;

calculate, for a particular substring of the plurality of substrings, a value relating to one or more documents, of the identified documents, that contain the particular substring;

determine that the calculated value for the particular substring satisfies a particular threshold associated with identifying compounds;

select, for a semantic unit, the particular substring from the plurality of substrings based on the calculated value for the particular substring satisfying the particular threshold; and

obtain a refined list of documents by refining the identified documents based on the semantic unit.

15. The system of claim 14, where the server is further to:

transmit the refined list of documents to a device of a user that submitted the query.

16. The system of claim 14,

where the server is further to:

receive the query via a network, and where, when identifying the documents, the server is to;

identify the documents relating to the query by matching individual search terms in the query to an index of a corpus that includes a collection of web documents.

17. The system of claim 14, where the server is further to:

discard, before selecting the particular substring, one or more substrings of the plurality of substrings that overlap with one or more other substrings of the plurality of substrings.

18. The system of claim 14, where the value is weighted based on a ranking defined by relevance of the identified documents.

19. The system of claim 14, where the server is further to:

provide information based on the refined list of documents.