Document length as a static relevance feature for ranking search results

• US 9,348,912 B2
• Filed: 09/10/2008
• Issued: 05/24/2016
• Est. Priority Date: 10/18/2007
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Tankovich, Vladimir, Meyerzon, Dmitriy, Taylor, Michael James
• Primary Examiner(s): Kim, Taelor
• Application Number:
  US12/207,910
• Publication Number:
  US 20090106235A1
• Time in Patent Office: 2,813 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

Embodiments are configured to provide information based on a user query. In an embodiment, a system includes a search component having a ranking component that can be used to rank search results as part of a query response. In one embodiment, the ranking component includes a ranking algorithm that can use the length of documents returned in response to a search query to rank search results.

Claims

1. A system for providing information comprising:

one or more processors;

one or more computer storage media storing computer executable instructions that when executed by the one or more processors provide;

a search component configured to locate a search result based on a query input;

a database component configured to store information associated with the query input including one or more ranking features, wherein the one or more ranking features are associated with a user action or user inaction associated with the search result which are collected with respect to the search result for a same query or a similar query previously received, and wherein one ranking feature of the one or more ranking features is associated with a normalized document length wherein the normalized document length is determined by dividing a length of a document to be ranked by an average length of a set of documents included in the search result, wherein the document to be ranked is included in the set of documents, wherein the length of the document corresponds to a number of words in the document; and

a ranking component configured to rank the search result based, at least in part, on a ranking function and the one or more ranking features, including an action-based feature, an inaction-based feature and a normalized document length feature, wherein the search component uses the rank of the search result when providing search results according to a ranking order.

2. The system of claim 1, wherein a transform function converts the normalized document length into a ranking value between zero and one.

3. The system of claim 2, wherein the transform function is defined as:

F(D)=D,D<

=1

F(D)=0.5+(3−

D)/4,1<

B<

=3

F(D)=2/(D+1),D>

3,

wherein, D represents the normalized document length and F(D) represents the ranking value.

4. The system of claim 1, wherein the ranking component uses one or more click-through parameters when ranking the search result, wherein the one or more click-through parameters further comprise one or more of the following:

a click parameter associated with a number of times that the search result has been clicked;

a skip parameter associated with a number of times that the search result has been skipped;

a first stream parameter corresponding to a union of query strings associated with a clicked search result; and

a second stream parameter corresponding to a union of query strings associated with a skipped search result.

5. The system of claim 4, wherein the search component is further configured to update the one or more click-through parameters including using information associated with how the search result was interacted with when updating the one or more of the click-through parameters.

6. The system of claim 5, wherein the search component is further configured to update the one or more click-through parameters, wherein the update of the one or more click-through parameters corresponds with a selected search result or a skipped search result.

7. The system of claim 1, wherein the wherein the one or more ranking features comprise one or more dynamic ranking features selected from a group consisting of body, title, author, generated title, an anchor text, and a URL, and one or more static ranking features selected from a group consisting of click distance, URL depth, file type, and language.

15. The system of claim 1, wherein the normalized document length is determined independently from a file type of one or more documents included in the search result.

8. A non-transitory computer-readable storage medium storing computer executable instructions that when executed by one or more processors provide a search engine configured to:

receive information associated with a query;

locate a search result associated with the query, wherein the search result includes one or more documents;

calculate a first input associated with a click parameter and the search result;

calculate a second input associated with a skip parameter and the search result;

calculate a third input associated with a normalized document length of the one or more documents included in the search result, wherein the normalized document length of the one or more documents is obtained by dividing a length of each document of the one or more documents by an average length of each document of the one or more documents included in the search result, wherein the length of the document corresponds to a number of words in the document;

wherein the length of the document corresponds to a number of words in the document;

store information associated with the query including one or more ranking features, wherein the one or more ranking features are associated with a user action or user inaction associated with the search result which are collected with respect to the search result for a same query or a similar query previously received, and wherein one ranking feature of the one or more ranking features is associated with a normalized document length, wherein the document to be ranked is included in the one or more documents;

ranking the search result based on a ranking determination using the ranking features, the first input, the second input, and the third input;

and provide the search result according to the ranking determination.

9. The non-transitory computer-readable storage medium of claim 8, further configured to

calculate a fourth input associated with a first stream parameter and the search result;

calculate a fifth input associated with a second stream parameter and the search result; and

rank the one or more documents included in the search result using at least four of the first input, the second input, the third input, the fourth input, and the fifth input.

10. The non-transitory computer-readable storage medium of claim 8, further configured to update a store with click parameter and skip parameter updates associated with received interactions with the one or more documents included in the search result.

11. The non-transitory computer-readable storage medium of claim 8, further configured to update a store with stream parameter updates associated with received interactions with the one or more documents included in the search result.

16. The non-transitory computer-readable storage medium of claim 8, wherein the normalized document length of the one or more documents included in the search result is determined independently from a file type of the one or more documents included in the search result.

17. The non-transitory computer-readable storage medium of claim 8, wherein the normalized document length of the one or more documents is obtained by dividing a length of each document of the one or more documents by an average length of each document of the one or more documents included in the search result.

12. A method of providing information comprising:

searching to locate a search result based on a query input;

storing information associated with the query input including one or more ranking features, wherein the one or more ranking features are associated with a user action or user inaction associated with the search result which are collected with respect to the search result for a same query or a similar query previously received, and wherein one ranking feature of the one or more ranking features is associated with a normalized document length wherein the normalized document length is determined by dividing a length of a document to be ranked by an average length of a set of documents included in the search result, wherein the document to be ranked is included in the set of documents, wherein the length of the document corresponds to a number of words in the document; and

ranking the search result based, at least in part, on a ranking function and the one or more ranking features, including an action-based feature, an inaction-based feature and a normalized document length feature, wherein the search component uses the rank of the search result when providing search results according to a ranking order.

13. The method of claim 12, further comprising:

determining a fourth input value associated with a text stream and a received selection of at least one of the one or more query candidates; and

ranking the one or more query candidates based in part on a scoring determination using a scoring function and one or more of the first input value, the second input value, the third input value and the fourth input value.

14. The method of claim 12, further comprising ranking the one or more query candidates according to a numerical order.

18. The method of claim 12, wherein the normalized document length of the at least one of the query candidates is determined independently from a file type of the at least one of the query candidates.

19. The method of claim 12, wherein the normalized document length of the at least one of the query candidates is obtained by dividing a length of the at least one of the query candidates by an average length of the one or more query candidates included in a result of the query.