Systems and methods for providing search query refinements

• US 10,223,439 B1
• Filed: 11/10/2016
• Issued: 03/05/2019
• Est. Priority Date: 09/30/2004
• Status: Active Grant

• Current Assignee: Google LLC (Alphabet Inc.)
• Original Assignee: Google LLC (Alphabet Inc.)
• Inventors: Baker, Steven, Haahr, Paul
• Primary Examiner(s): Trujillo, James
• Assistant Examiner(s): Cheung, Hubert
• Application Number:
  US15/348,314
• Time in Patent Office: 845 Days
• CPC Class Codes: G06F 16/3322 using system suggestions G06F16/3325 takes precedence
• CPC Class Codes: G06F 16/3325 Reformulation based on results of preceding query
• CPC Class Codes: G06F 16/93 Document management systems
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation

Abstract

A system and method for generating query refinement suggestions may include collecting refinement data for at least one received source query. The collected refinement data is then clustered to form at least one cluster. At least one potential refinement query suggestion is identified from the refinement data within the at least one cluster.

Claims

1. A computer implemented method, comprising:

receiving, by one or more devices, an audible search query from a voice recognition system;

identifying, by the one or more devices, a plurality of subsequent search queries that were previously received after the audible search query;

the plurality of subsequent search queries being refinements of the audible search query;

identifying, by the one or more devices, a first quantity of terms that occur in a first set of documents associated with a first subsequent search query of the plurality of subsequent search queries;

the first quantity of terms being based on the audible search query;

identifying a respective weight associated with each term of the first quantity of terms that occur in the first set of documents;

generating, based on identifying the respective weight associated with each term of the first quantity of terms, a term vector for each document in the first set of documents;

identifying, by the one or more devices, a second quantity of terms that occur in a second set of documents associated with a second subsequent search query of the plurality of subsequent search queries;

the second quantity of terms being based on the audible search query;

identifying a respective weight associated with each term of the second quantity of terms that occur in the second set of documents;

generating, based on identifying the respective weight associated with each term of the second quantity of terms, a term vector for each document in the second set of documents;

selecting, by the one or more devices and based on the identified first quantity of terms and the identified second quantity of terms, the first subsequent search query or the second subsequent search query; and

providing output through a speaker, by the one or more devices, the output being the selected first subsequent search query or the selected second subsequent search query as a refinement of the audible search query.

2. The computer implemented method of claim 1 further comprising:

normalizing the generated term vector for each document in the first set of documents; and

normalizing the generated term vector for each document in the second set of documents.

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

creating, based on normalizing the generated term vector for each document in the first set of documents and normalizing the generated term vector for each document in the second set of documents, a composite term vector.

4. The computer implemented method of claim 3 where, when selecting the first subsequent search query or the second subsequent search query, the computer implemented method includes:

selecting the first subsequent search query or the second subsequent search query using the composite term vector.

5. The computer implemented method of claim 3 further comprising:

modifying the composite term vector to exclude at least one of the first quantity of terms or the second quantity of terms that does not satisfy a particular threshold value.

6. The computer implemented method of claim 1 where, when selecting the first subsequent search query or the second subsequent search query, the computer implemented method includes:

selecting the first subsequent search query or the second subsequent search query based on a frequency of occurrence associated with the first subsequent search query and a frequency of occurrence associated with the second subsequent search query.

7. A system comprising:

one or more memory devices to store instructions;

one or more processors to execute the instructions to;

receive by the one or more devices an audible search query from a voice recognition system;

identify, by the one or more devices, a plurality of subsequent audible search queries that were previously received after the audible search query;

the plurality of subsequent search queries being refinements of the audible search query;

identify a first quantity of terms that occur in a first set of documents associated with a first subsequent search query of the plurality of subsequent search queries,

the first quantity of terms being based on the audible search query;

identify a respective weight associated with each term of the first quantity of terms that occur in the first set of documents;

generate, based on identifying the respective weight associated with each term of the first quantity of terms, a term vector for each document in the first set of documents;

identify a second quantity of terms that occur in a second set of documents associated with a second subsequent search query of the plurality of subsequent search queries,

the second quantity of terms being based on the audible search query;

identify a respective weight associated with each term of the second quantity of terms that occur in the second set of documents;

generate, based on identifying the respective weight associated with each term of the second quantity of terms, a term vector for each document in the second set of documents;

select, based on the identified first quantity of terms and the identified second quantity of terms, the first subsequent search query or the second subsequent search query; and

provide output through a speaker, by the one or more devices, the output being the selected first subsequent search query or the selected second subsequent search query as a refinement of the audible search query.

8. The system of claim 7 where the one or more processors are further to:

normalize the generated term vector for each document in the first set of documents; and

normalize the generated term vector for each document in the second set of documents.

9. The system of claim 8 where the one or more processors are further to:

create, based on normalizing the generated term vector for each document in the first set of documents and normalizing the generated term vector for each document in the second set of documents, a composite term vector.

10. The system of claim 9 where, when selecting the first subsequent search query or the second subsequent search query, the one or more processors are further to:

select the first subsequent search query or the second subsequent search query using the composite term vector.

11. The system of claim 9 where the one or more processors are further to:

modify the composite term vector to exclude at least one of the first quantity of terms or the second quantity of terms that does not satisfy a particular threshold value.

12. The system of claim 7, where, when selecting the first subsequent search query or the second subsequent search query, the one or more processors are further to:

select the first subsequent search query or the second subsequent search query based on a frequency of occurrence associated with the first subsequent search query and a frequency of occurrence associated with the second subsequent search query.

13. A system comprising:

one or more memory devices storing instructions; and

one or more processors to execute the instructions to;

identify a plurality of audible subsequent search queries that were received from a voice recognition system after a particular audible search query received from the voice recognition system, as refinements of the particular audible search query;

identify a first quantity of terms that occur in a first set of documents that is associated with a first audible subsequent search query received from a voice recognition system, of the plurality of audible subsequent search queries;

identify a first quantity of terms that occur in the first set of documents associated with the first audible subsequent search query of the plurality of subsequent search queries;

identify a respective first weight associated with each term of the first quantity of terms that occur in the first set of documents;

identify a second set of documents that is associated with a second audible subsequent search query received from a voice recognition system, of the plurality of audible subsequent search queries;

identify a second quantity of terms that occur in the second set of documents associated with the second audible subsequent search query of the plurality of subsequent search queries;

identify a respective second weight associated with each term of the second quantity of terms that occur in the second set of documents;

generate a first term vector for the first audible subsequent search query based on the identified first set of documents and based on the respective first weight, and a second term vector for the second audible subsequent search query based on the identified second set of documents and based on the respective second weight;

cluster, based on the generated first term vector and the generated second term vector, the first audible subsequent search query into a first cluster and the second audible subsequent search query into a different, second cluster;

select a particular audible subsequent search query from one or more of the first cluster or the different, second cluster;

receive an indication that a client device has submitted an audible search query that includes the particular search query; and

provide, to the client device, output through a speaker, the selected particular audible subsequent search query from the one or more of the first cluster or the different, second cluster, as a refinement of the particular audible search query submitted by the client device.