Using domain intent to provide more search results that correspond to a domain

• US 8,504,561 B2
• Filed: 09/02/2011
• Issued: 08/06/2013
• Est. Priority Date: 09/02/2011
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Hoad, Timothy C., Vijaywargi, Deepak, Saraf, Yatharth
• Primary Examiner(s): DANG, THANH HA T
• Application Number:
  US13/224,961
• Publication Number:
  US 20130060761A1
• Time in Patent Office: 704 Days
• Field of Search: 707/723, 707/750, 707/766, 707/769
• US Class Current: 707/723
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9538 Presentation of query results

Abstract

Techniques are described herein for using intent to access a domain (i.e., domain intent) to provide more search results that correspond to the domain. For example, a rule may specify a maximum number of search results that are allowed to be provided from a domain (or a host that corresponds to the domain) in response to a search query. Each search query may include any number of ngrams. An ngram is a subsequence of elements in a sequence (e.g., a search query). An intent to access a domain may be determined based on one or more of the ngrams in a search query. A number of search results that correspond to a domain may be increased to be greater than the maximum number based on one or more of the ngrams that are included in the search query being associated with the intent to access the domain.

Claims

1. A method comprising:

for each ngram of a plurality of ngrams, determining a click entropy that represents an extent to which a plurality of clicks that occurs with respect to a plurality of search queries that include the ngram is diversified among a plurality of domains based on a proportion of the plurality of clicks that corresponds to each domain of the plurality of domains;

comparing the click entropy for each ngram to an entropy threshold to determine a subset of the plurality of ngrams such that the click entropy of each ngram in the subset is less than the entropy threshold;

determining that each ngram in the subset is associated with an intent to access a respective designated domain of the plurality of domains, the designated domain for each ngram in the subset corresponding to a relatively greater proportion of the plurality of clicks that occurs with respect to the plurality of search queries that include that ngram than others of the plurality of domains; and

increasing a number of search results that correspond to the designated domain for a specified ngram from the subset that are to be provided in response to receipt of a search query that includes the specified ngram to be greater than a specified maximum number, using at least one processor, based on the specified ngram being associated with the intent to access the designated domain for the specified ngram.

2. The method of claim 1, wherein the click entropy for each ngram is based on a summation of values for that ngram, each value being based on a logarithmic function of the respective proportion of the plurality of clicks that occurs with respect to the plurality of search queries that include the ngram that corresponds to the respective domain.

3. The method of claim 1, wherein increasing the number of the search results that correspond to the designated domain for the specified ngram that are to be provided in response to receipt of the search query comprises:

increasing the number of the search results that correspond to the designated domain for the specified ngram that are to be provided in response to receipt of the search query to be a specified number that is selected based on the click entropy for the specified ngram.

4. The method of claim 1, wherein increasing the number of the search results that correspond to the designated domain for the specified ngram that are to be provided in response to receipt of the search query comprises:

limiting the number of the search results that correspond to the designated domain for the specified ngram that are to be provided in response to receipt of the search query to be no greater than a threshold number that is based on the click entropy for the specified ngram.

5. The method of claim 1, further comprising:

grouping the search results that correspond to the designated domain for the specified ngram that are to be provided in response to receipt of the search query to be consecutively ranked in a plurality of search results that is to be presented to a user that is associated with the search query.

6. The method of claim 5, further comprising:

determining that the number of the search results that correspond to the designated domain for the specified ngram that are to be provided in response to receipt of the search query is greater than or equal to a threshold number;

wherein grouping the search results that correspond to the designated domain for the specified ngram that are to be provided in response to receipt of the search query comprises;

grouping the search results that correspond to the designated domain for the specified ngram that are to be provided in response to receipt of the search query in response to determining that the number of the search results that correspond to the designated domain for the specified ngram that are to be provided in response to receipt of the search query is greater than or equal to a threshold number.

7. The method of claim 1, further comprising:

grouping the search results that correspond to the designated domain for the specified ngram that are to be provided in response to receipt of the search query to provide a group of consecutively listed search results, a highest ranking search result in the group being listed above other search results in the group, the other search results in the group being indented with respect to the highest ranking search result in the group.

8. A method comprising:

for each ngram of a plurality of ngrams, determining a domain click value that represents a proportion of a plurality of clicks, which occur with respect to a plurality of search queries that include the ngram, that corresponds to a designated domain;

comparing the domain click value for each ngram to a threshold value to determine a subset of the plurality of ngrams such that the domain click value of each ngram in the subset is greater than the threshold value;

determining that each ngram in the subset is associated with an intent to access the designated domain based on the domain click value for each ngram in the subset being greater than the threshold value; and

increasing a number of search results that correspond to the designated domain that are to be provided in response to receipt of a search query that includes a specified ngram from the subset to be greater than a specified maximum number, using at least one processor, based on the specified ngram being associated with the intent to access the designated domain.

9. The method of claim 8, wherein determining the domain click value for each ngram comprises:

for each ngram, determining a domain click value that represents a proportion of a plurality of clicks, which occur with respect to a plurality of search queries that include the ngram, that corresponds to each of a plurality of domains, the plurality of domains including the designated domain;

wherein the method further comprises;

for each ngram, summing intermediate values that are based on the respective domain click values for the ngram to provide a click entropy for that ngram; and

comparing the click entropy for each ngram to an entropy threshold to determine a subset of the click entropies, each click entropy in the subset having an absolute value that is less than the entropy threshold;

wherein increasing the number of the search results comprises;

increasing the number of the search results further based on the click entropy of the specified ngram being included in the subset of the click entropies.

10. The method of claim 9, further comprising:

for each ngram, performing a logarithmic operation with respect to each domain click value that corresponds to the ngram to provide the respective intermediate value.

11. The method of claim 8, wherein increasing the number of the search results that correspond to the designated domain that are to be provided in response to receipt of the search query comprises:

increasing the number of the search results that correspond to the designated domain that are to be provided in response to receipt of the search query to be a specified number that is selected based on the domain click value for the specified ngram.

12. The method of claim 8, wherein increasing the number of the search results that correspond to the designated domain that are to be provided in response to receipt of the search query comprises:

limiting the number of the search results that correspond to the designated domain that are to be provided in response to receipt of the search query to be no greater than a threshold number that is based on the domain click value for the specified ngram.

13. The method of claim 8, further comprising:

grouping the search results that correspond to the designated domain that are to be provided in response to receipt of the search query to be consecutively ranked in a plurality of search results that is to be presented to a user that is associated with the search query.

14. The method of claim 13, further comprising:

determining that the number of the search results that correspond to the designated domain that are to be provided in response to receipt of the search query is greater than or equal to a threshold number;

wherein grouping the search results that correspond to the designated domain that are to be provided in response to receipt of the search query comprises;

grouping the search results that correspond to the designated domain that are to be provided in response to receipt of the search query in response to determining that the number of the search results that correspond to the designated domain that are to be provided in response to receipt of the search query is greater than or equal to the threshold number.

15. The method of claim 8, further comprising:

grouping the search results that correspond to the designated domain that are to be provided in response to receipt of the search query to provide a group of consecutively listed search results, a highest ranking search result in the group being listed above other search results in the group, the other search results in the group being indented with respect to the highest ranking search result in the group.

16. A method comprising:

establishing a rule that specifies a maximum number of search results that are to be provided from each of a plurality of domains in response to an arbitrary search query;

determining that at least one of a click entropy associated with an ngram is less than an entropy threshold or a domain click value associated with the ngram is greater than a threshold value,

the ngram identified to be associated with an intent to access a specified domain of the plurality of domains,

the click entropy representing an extent to which a plurality of clicks that occurs with respect to a plurality of search queries that include the ngram is diversified among the plurality of domains,

the domain click value representing a proportion of the plurality of clicks that occurs with respect to the plurality of search queries that include the ngram that corresponds to the specified domain; and

dynamically increasing a number of search results that are to be provided from the specified domain in response to a specified search query to be greater than the maximum number, using at least one processor, based on the specified search query including the ngram and further based on at least one of the click entropy being less than the entropy threshold or the domain click value being greater than the threshold value.

17. The method of claim 16, further comprising:

re-ranking one or more of the search results that are to be provided from the specified domain such that the search results that are to be provided from the specified domain are consecutively ranked in a plurality of search results that is to be presented to a user that is associated with the specified search query.

18. The method of claim 17, further comprising:

determining that the number of the search results that are to be provided from the specified domain in response to the specified search query is greater than or equal to a threshold number;

wherein re-ranking the one or more of the search results that are to be provided from the specified domain comprises;

re-ranking the one or more of the search results that are to be provided from the specified domain based on the number of the search results that are to be provided from the specified domain in response to the specified search query being greater than or equal to the threshold number.

19. The method of claim 16, wherein dynamically increasing the number of the search results that are to be provided from the specified domain in response to the specified search query comprises:

limiting the number of the search results that are to be provided from the specified domain in response to the specified search query to be no greater than a threshold number that is based on the click entropy associated with the ngram.

20. The method of claim 16, wherein dynamically increasing the number of the search results that are to be provided from the specified domain in response to the specified search query comprises:

limiting the number of the search results that are to be provided from the specified domain in response to the specified search query to be no greater than a threshold number that is based on the domain click value associated with the ngram.