Determining relevance of a document to a query based on spans of query terms

• US 20070094234A1
• Filed: 10/26/2005
• Published: 04/26/2007
• Est. Priority Date: 10/26/2005
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Ma, Wei-Ying, Wen, Ji-Rong, Song, Ruihua
• Granted Patent: US 7,480,652 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/334 Query execution G06F16/335 takes precedence
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: Y10S 707/99931 Database or file accessing
• CPC Class Codes: Y10S 707/99932 Access augmentation or optimizing
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching
• CPC Class Codes: Y10S 707/99934 Query formulation, input preparation, or translation
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access

Abstract

A relevance system determines the relevance of a query term to a document based on spans within the document that contain the query term. The relevance system aggregates the relevance of the query terms into an overall relevance for the document. For each query term, the relevance system calculates a span relevance for each span that contains that query term. The relevance system then aggregates the span relevances for a query term into a query term relevance for that document. The relevance system may aggregate the query term relevances into a document relevance.

Claims

1. A method in a computer system for determining relevance of a query term to a document, the method comprising:

identifying spans of query terms within the document;

calculating a relevance for the query term for each span that includes the query term; and

aggregating the calculated relevance for the query term for each span into the relevance of the query term to the document.

2. The method of claim 1 wherein the identifying of spans includes identifying spans of query terms wherein a span includes no repeated occurrences of a query term and wherein a distance between adjacent query terms within a span is less than a threshold distance.

3. The method of claim 2 wherein a span includes a repeated occurrence of a query term that is closest to a query term of the span.

4. The method of claim 1 wherein the relevance for the query term for a span is calculated based on the following:

$f<br><br>\left(t,{\mathrm{espan}}_i\right)=\frac{n_i^y}{{\mathrm{Width}<br><br>\left({\mathrm{espan}}_i\right)}^x}$

where t is a query term, espan_i is a span that contains t, n_i is the number of query terms that occur in espan_i, Width(espan_i) is the span width of espan_i, x is an exponent that is used to control the influence of the span width, and y is an exponent that is used to control the influence of the number of query terms in the span.

5. The method of claim 1 wherein the aggregating includes summing the calculated relevance for the query term for each span.

6. The method of claim 1 including replacing a term frequency with the aggregated relevance in an equation for calculating relevance of a document to a query.

7. The method of claim 6 wherein the equation before replacing is based on term frequency and inverse document frequency.

8. The method of claim 6 wherein term frequency is not used to calculate relevance.

9. The method of claim 1 wherein the document is a web page.

10. A computer-readable medium containing instructions for controlling a computer system to identify spans of query terms within a document, by a method comprising:

identifying query terms within the document; and

identifying sequences of query terms wherein a sequence includes no repeated occurrences of a query term and wherein at least one sequence includes at least three query terms wherein a sequence is a span of query terms.

11. The computer-readable medium of claim 10 wherein a distance between adjacent query terms within a sequence is less than a threshold distance.

12. The computer-readable medium of claim 11 wherein the distance between two query terms is based on the number of non-query terms between the two query terms.

13. The computer-readable medium of claim 10 wherein when a sequence of query terms is adjacent to two occurrences of the same query term, including within the sequence the query term that is closest to a query term of the sequence.

14. The computer-readable medium of claim 10 including calculating a relevance of query terms based on the following:

$f<br><br>\left(t,{\mathrm{espan}}_i\right)=\frac{n_i^y}{{\mathrm{Width}<br><br>\left({\mathrm{espan}}_i\right)}^x}$

where t is a query term, espan_i is a span that contains t, n_i is the number of query terms that occur in espan_i, Width(espan_i) is the span width of espan_i, x is an exponent that is used to control the influence of the span width, and y is an exponent that is used to control the influence of the number of query terms in the span.

15. A method in a computer system for calculating relevance of a document to a query of query terms, the method comprising:

identifying spans of query terms within the document;

calculating a relevance contribution of each query term based on the identified spans; and

determining relevance of the document to the query based on the calculated relevance contributions without using term frequency.

16. The method of claim 15 wherein the identifying of spans includes identifying sequences of query terms wherein a sequence includes no repeated occurrences of a query term and wherein at least one sequence includes at least three query terms.

17. The method of claim 16 wherein a distance between adjacent query terms within a sequence is less than a threshold distance.

18. The method of claim 15 wherein the calculating of the relevance contribution of a query term includes calculating relevance of the query term resulting from a span based on the following:

$f<br><br>\left(t,{\mathrm{espan}}_i\right)=\frac{n_i^y}{\mathrm{Width}<br><br>\text{\hspace{1em}<br><br>}<br><br>{\left({\mathrm{espan}}_i\right)}^x}$

where t is a query term, espan_i is a span that contains t, n_i is the number of query terms that occur in espan_i, Width(espan_i) is the span width of espan_i, x is an exponent that is used to control the influence of the span width, and y is an exponent that is used to control the influence of the number of query terms in the span.

19. The method of claim 15 wherein the determining of relevance is based on an equation that uses inverse document frequency.

20. The method of claim 19 wherein term frequency within the equation is replaced with the calculated relevance contribution.