Classification of search queries

• US 9,767,182 B1
• Filed: 10/28/2016
• Issued: 09/19/2017
• Est. Priority Date: 10/28/2016
• Status: Active Grant

• Current Assignee: Searchmetrics GmbH
• Original Assignee: Searchmetrics GmbH
• Inventors: Thakur, Abhishek, Tober, Marcus
• Primary Examiner(s): Holmes, Michael B
• Application Number:
  US15/337,163
• Time in Patent Office: 326 Days
• CPC Class Codes: G06F 16/35 Clustering; Classification
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06N 3/045 Combinations of networks

Abstract

A computer-implemented method of classifying a search query in a network comprises: classifying a plurality of search queries into categories, comprising: applying predetermined rules to each of the plurality of search queries, wherein the predetermined rules are indicative of the categories and each of the plurality of search queries is associated with search results in the network; determining, for each of the plurality of search queries, similarity values indicating similarity to each of the categories based on the applied predetermined rules; and training a machine learning module, comprising: applying the machine learning module to a plurality of training sets to a plurality of training sets, wherein each of the plurality of training sets is based on one of the plurality of classified search queries and at least one of the respective one or more similarity values, a corresponding system, computing device and non-transitory computer-readable storage medium.

Claims

1. A computer-implemented method of classifying a search query in a network, the method comprising:

obtaining, from the network, at least one search result for each of a plurality of search queries, wherein the plurality of search queries includes at least one unseen search query;

modifying at least one of the plurality of search queries comprising at least one unseen search query;

with a processing module, executing on one or more computing devices, automatically;

classifying the plurality of search queries into one or more categories, the classifying comprising;

applying one or more predetermined rules to each of the plurality of search queries, wherein the one or more predetermined rules are indicative of the one or more categories and each of the plurality of search queries is associated with one or more search results in the network;

determining, for each of the plurality of search queries, one or more similarity values indicating similarity to each of the one or more categories based on the applied one or more predetermined rules;

with a training module, executing on one or more computing devices, automatically;

training a machine learning module using the classified search queries including the at least one unseen search query, the training comprising;

modifying at least one of the plurality of classified search queries in order to change a plurality of training sets to be applied to the training module;

obtaining a training set from the plurality of training sets;

obtaining each predetermined rule from the one or more predetermined rules;

applying the each predetermined rule to the training set to determine a similarity of the training set to the one or more categories;

evaluating the similarities for the training set to the one or more categories to classify the training set, wherein each training set is classified to the one or more categories;

converting each of the plurality of training sets and associated similarity values into a number or a vector;

determining whether the each of the plurality of training sets is within a range suitable for the machine learning module;

applying the machine learning module to the plurality of training sets, wherein the each of the plurality of training sets is based on one of the plurality of classified search queries and at least one of the respective one or more similarity values

receiving the at least one unseen search query; and

classifying the at least one unseen search query into a category of the one or more categories by applying the trained machine learning module.

2. The method of claim 1, wherein:

the two or more categories comprise at least one of;

one or more main categories; and

one or more user intent categories.

3. The method of claim 2, wherein:

the one or more user intent categories comprise at least one of;

an informational category;

a navigational category; and

a transactional category.

4. The method of claim 1, wherein:

the one or more predetermined rules comprise at least one of;

one or more informational rules;

one or more navigational rules; and

one or more transactional rules.

5. The method of claim 4, wherein:

the one or more informational rules comprise at least one of;

query comprises one or more question words;

query comprises one or more natural language terms;

query comprises one or more informational terms;

query is subsequent query;

query resulted in viewing of more than one search result;

query comprises more than two terms; and

query that does not match navigational rule or transactional rule;

the one or more navigational rules comprise at least one of;

query comprises name, company name, business name, organization name or personal name;

query comprises one or more domain suffixes;

query with “

web”

as the source;

query comprises less than three terms; and

query resulted in viewing of first search engine results page;

or

the one or more transactional rules comprise at least one of;

query comprises one or more terms relating to at least one of movies, songs, lyrics, recipes, images, humor and porn;

query comprises one or more terms relating to obtainment;

query comprises one or more terms relating to download;

query relates to at least one of an image collection, and audio collection or a video collection;

query with “

audio”

, “

images”

, or “

video”

as the source;

query comprises one or more terms relating to entertainment;

query comprises one or more terms relating to at least one of interaction, buying and chatting;

query comprises one or more terms relating to movies, songs, lyrics, images; and

query comprises at least one of multimedia file extension and compression file extension.

6. The method of claim 1, wherein:

the two or more categories comprise at least one of;

one or more sub categories; and

one or more buying cycle categories.

7. The method of claim 6, wherein:

the one or more buying cycle categories comprise at least one of;

an awareness category;

a consideration category;

a purchase category; and

a retention category.

8. The method of claim 1, wherein:

the one or more search results of each of the plurality of search queries comprise at least one of;

one or more titles of one or more resources;

one or more identifiers to one or more resources; and

one or more integrations supplemented to the one or more search results.

9. The method of claim 1, wherein:

the at least one of the respective one or more similarity values is a highest similarity value within the one or more similarity values.

10. The method of claim 1, wherein:

the machine learning module comprises at least one of a neural network, deep neural network, convolutional neural network, linear model and generalized boosted model.

11. The method of claim 1, further comprising:

with a search results module, executing on one or more computing devices, automatically;

obtaining, from the network, the one or more search results for each of the plurality of search queries; and

storing the one or more search results of each of the plurality of search queries.

12. The method of claim 1, further comprising:

with the processing module, prior to classifying the plurality of search queries into the two or more categories, automatically;

modifying at least one of the plurality of search queries by deleting a term, word, filler word, nuisance word or stopword;

modifying at least one of the plurality of search queries by replacing a term with a modified term or basic term corresponding to the term;

processing one or more titles of one or more resources;

or

processing one or more identifiers to one or more resources.

13. The method of claim 1, wherein:

each of the plurality of training sets comprises;

a plurality of numbers;

or

a vector.

14. The method of claim 1, further comprising:

with the training module, prior to training the machine learning module, automatically;

converting the plurality of classified search queries and the at least one of the respective one or more similarity values into the plurality of training sets.

15. A system for classifying a search query, the system comprising:

one or more processors, when executing on one or more computing devices;

being suitable for performing the method of claim 1.

19. A non-transitory computer-readable storage medium comprising instructions causing a system to perform operations for classifying a search query, the operations being suitable for performing the method of claim 1.

16. A computing device for classifying a search query, the computing device comprising:

one or more processors, configured to perform operations; and

a memory, coupled to the one or more processors and comprising instructions to cause, when executing on the one or more processors, the computing device to perform operations, comprising;

obtaining, from the network, at least one search result for each of a plurality of search queries, wherein the plurality of search queries includes at least one unseen search query;

modifying at least one of a plurality of search queries comprising at least one unseen search query;

classifying the plurality of search queries into one or more categories, the classifying comprising;

applying one or more predetermined rules to each of the plurality of search queries, wherein the one or more predetermined rules are indicative of the one or more categories and each of the plurality of search queries is associated with one or more search results in the network;

determining, for each of the plurality of search queries, one or more similarity values indicating similarity to each of the one or more categories based on the applied one or more predetermined rules;

training a machine learning module using the classified search queries including the at least one unseen search query, the training comprising;

modifying at least one of the plurality of classified search queries in order to change a plurality of training sets to be applied to the training module;

obtaining a training set from the plurality of training sets;

obtaining each predetermined rule from the one or more predetermined rules;

applying the each predetermined rule to the training set to determine a similarity of the training set to the one or more categories;

evaluating the similarities for the training set to the one or more categories to classify the training set, wherein each training set is classified to the one or more categories;

converting each of the plurality of training sets and associated similarity values into a number or vector;

determining whether the each of the plurality of training sets is within a range suitable for the machine learning module;

applying the machine learning module to a plurality of training sets, wherein each of the plurality of training sets is based on one of the plurality of classified search queries and at least one of the respective one or more similarity values;

receiving the at least one unseen search query; and

classifying the at least one unseen search query into a category of the one or more categories by applying the trained machine learning module.

17. The computing device of claim 16, wherein:

the two or more categories comprise at least one of;

one or more main categories; and

one or more user intent categories.

18. The computing device of claim 16, wherein:

the one or more user intent categories comprise at least one of;

an informational category;

a navigational category; and

a transactional category.