DOMAIN-BASED RANKING IN DOCUMENT SEARCH

• US 20100228743A1
• Filed: 03/03/2009
• Published: 09/09/2010
• Est. Priority Date: 03/03/2009
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Adamson, Daniel, Rappaport, Alain Thierry
• Granted Patent: US 9,836,538 B2
• Time in Patent Office: Days
• US Class Current: 707/750
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques

Abstract

In one example, documents that are examined by a search process may be scored in a manner that is specific to a domain. A domain may be a substantive area, such as medicine, sports, etc. Different scoring approaches that take aspects of the domain into account may be applied to the documents, thereby producing different scores than might have been produced by a simple comparison of the terms in the query with the terms in the documents. These domain-based approaches may take a query into account in scoring the documents, or may be query-independent. Each approach may be implemented by a scorer. The combined output of the scorers may be used to generate a score for each document. Documents then may be ranked based on the scores, and search results may be provided.

Claims

1. A computer-readable storage medium that stores executable instructions that, when executed by a computer, cause the computer to perform acts comprising:

receiving a query;

determining scores of documents in a corpus based on a comparison of terms in said query with terms in said documents;

identifying a set of scorers, from among a plurality of scorers, that will contribute to scoring of said documents, each of said plurality of scorers evaluating said documents based on aspects of a substantive domain;

adjusting a score of each of said documents based on a contribution from each scorer in said set of scorers, each document'"'"'s score after said adjusting being referred to as an adjusted score; and

presenting a set of search results to a user based on said adjusted scores.

2. The computer-readable storage medium of claim 1, wherein said acts further comprise:

identifying a set of concepts associated with terms in said query; and

increasing a score of a given one of said documents based on how many concepts in said set of concepts are in said given one of said documents.

3. The computer-readable storage medium of claim 2, further comprising:

reducing said score of a given one of said documents based on said given one of said documents having an amount of concepts from said set of concepts in excess of a number.

4. The computer-readable storage medium of claim 1, wherein said acts further comprise:

identifying a set of concepts associated with terms in said query;

and either;

decreasing a score of a given one of said documents based on how many concepts in said set of concepts are not in said given one of said documents;

or

decreasing a score of said given one of said documents based on how many concepts in said one of said documents are not in said set of concepts.

5. The computer-readable storage medium of claim 1, further comprising:

identifying a first set of concepts in said query;

identifying a second set of concepts in a summary of a given one of said documents;

determining that said first set has a defined level of similarity to said second set; and

based on said first set having said defined level of similarity to said second set, increasing a score of said given one of said documents.

6. The computer-readable storage medium of claim 1, wherein a first one of said scorers evaluates said documents'"'"' without regard to said query.

7. The computer-readable storage medium of claim 1, further comprising:

determining a number of concepts from said domain that appear in a given one of said documents;

determining that said number falls within a range; and

modifying a score of said given one of said documents based on said number'"'"'s falling within said range.

8. The computer-readable storage medium of claim 1, further comprising:

determining a number of concepts from said domain that appear in a given one of said documents; and

modifying a score of said given one of said documents by an amount that is based on said number.

9. The computer-readable storage medium of claim 1, further comprising:

determining that a given one of said documents has concept, from said domain, that has a level of popularity across said documents; and

based on said concept being in said given one of said documents, modifying a score of said given one of said documents.

10. The computer-readable storage medium of claim 1, further comprising:

calling a function in each of said scorers to determine which of said scorers will contribute to scoring of said documents, wherein each of said scorers implements its own version of said function.

11. A system that responds to a document search request, the system comprising:

a processor;

a scoring system that executes on said processor, the scoring system comprising;

a term-based scorer that performs a comparison of terms in a query to terms in a plurality of documents and that scores said documents based on said comparison;

a plurality of domain-based scorers that score said documents based on aspects of said documents that relate to a substantive domain; and

an aggregator that combines contributions from said term-based scorer and said plurality of domain-based scorers into scores for each of the documents; and

a presenter that presents, to a user, results of a search that are based on said scores.

12. The system of claim 11, wherein a first one of said plurality of domain-based scorers identifies a set of concepts associated with terms in said query and contributes to an increase in score of a given one of said documents based on how many concepts in said set of concepts are in said given one of said documents.

13. The system of claim 11, wherein a first one of said plurality of domain-based scorers identifies a first set of concepts in said query, identifies a second set of concepts in a summary of a given one of said documents, wherein said first one of said domain-based scorers contributes to an increase in a score of said given one of said documents based on said first set being identical to said second set.

14. The system of claim 11, wherein a first one of said plurality of domain-based scorers determines that a number of concepts from said domain that appear in a given one of said documents falls within a range, and contributes to a modification in score of said given one of said documents based on said number falling within said range.

15. The system of claim 11, wherein a first one of said plurality of domain-based scorers determines a number of concepts of said domain that appear in a given one of said documents and contributes to a modification in score of said given one of said documents, said increase being in an amount that is based on said number.

16. The system of claim 11, wherein a first one of said plurality of domain-based scorers determines that a given one of said documents has a concept that has a level of popularity across said documents, and, based on said concept being in said given one of said documents, contributes to a modification in score of said given one of said documents.

17. The system of claim 11, wherein each of said scorers exposes a callable function that, when called, indicates whether said scorer will contribute to scoring of said documents.

18. The system of claim 11, wherein said domain comprises a medical domain that includes concepts that describe either conditions of a human body or medical treatments of said human body.

19. A method of responding to a search query, the method comprising:

using a processor to perform acts comprising;

receiving a query;

calculating scores for a plurality of documents by comparing terms in said query with terms in said documents;

calling a function exposed by each of a plurality of domain-based scorers to determine which of said domain-based scorers will contribute to scoring of said documents, each of said domain-based scorers calculating a score based on features of said documents or of said query that are specific to a substantive field of knowledge;

including, on a list, those scorers that indicate, through said function, that they will contribute to scoring of said documents;

using those ones of said scorers that are on said list to adjust said scores, whereby adjusted scores of said documents are created;

creating a set of search results based on the adjusted scores of said documents; and

presenting said search results to a user.

20. The method of claim 19, wherein each of the domain-based scorers that are on said list contributes a score for each of said documents, and wherein said acts further comprise:

calculating said adjusted score for each document by calculating a product of;

(a) a score for that document that is calculated by comparing terms in said query with terms in the document, and (b) the scores contributed by each of the domain-based scorers that are on said list.