Assigning relevance weights based on temporal dynamics

Patent metadata

Field	Value
Publication number	US-10353967-B2
Application number	US-48859409-A
Country	US
Kind code	B2
Filing date	Jun 22, 2009
Priority date	Jun 22, 2009
Publication date	Jul 16, 2019
Grant date	Jul 16, 2019

Abstract

A system described herein includes a receiver component that receives a first dataset, wherein the first dataset comprises temporal dynamics pertaining to a document that is accessible by a search engine, wherein the temporal dynamics comprise an identity of a term corresponding to the document and an indication that the term has been subject to change over time. The system also includes a weight assignor component that assigns a relevance weight to the document based at least in part upon the temporal dynamics pertaining to the document, wherein the relevance weight is utilized by the search engine to assign a ranking to the document with respect to at least one other document when the search engine retrieves the document.

First claim

What is claimed is:

1. A method performed by at least one computing device that executes a search engine, the method comprising:
receiving, via an input interface of the search engine, a query that comprises a term;
classifying the query as being a navigational query rather than a seeking query;
assigning a relevance weight to a document based upon:
a term classification assigned to the term in the document, the term classification being representative of a length of time that the term has been present in the document over a window of time; and
the query being classified as a navigational query rather than a seeking query,

wherein the relevance weight is indicative of relevance of the document with respect to the query; and
causing a ranked list of search results to be presented on a display responsive to assigning the relevance weight to the document, the ranked list of search results comprises the document, wherein a position of the document in the ranked list of search results is based upon the relevance weight assigned to the document.

2. The method of claim 1 , further comprising:
assigning a second weight to the document that is independent of the query; and
assigning the relevance weight to the document based upon the second weight.

3. The method of claim 1 , further comprising:
responsive to receiving the query, retrieving first data, the first data includes a first mixture component and a second mixture component, the first mixture component comprises a first plurality of terms that have a rate of change in the document above a first threshold, the second mixture component comprises a second plurality of terms that have a rate of change in the document below the first threshold, the term included in the first plurality of terms of the first mixture component; and
assigning the relevance weight to the document based upon the term being included in the first plurality of terms of the first mixture component.

4. The method of claim 3 , wherein the first data includes a function that represents a continuous representation of the document over time with respect to rates of changes of terms in the document.

5. The method of claim 1 , wherein the term is in a body of the document, the method further comprising assigning the relevance weight to the document based upon the term being included in the body of the document.

6. The method of claim 1 , wherein the term is in anchor text of the document, the method further comprising assigning the relevance weight to the document based upon the term being in the anchor text of the document.

7. The method of claim 1 , further comprising assigning the relevance weight to the document based upon identities of portions of the document that have altered over time.

8. The method of claim 1 , further comprising assigning the relevance weight to the document based upon a change of position of the term in the document over time.

9. The method of claim 1 , the method further comprising:
generating a first language model for the document, the first language model comprises terms that occur in instances of the document above a threshold frequency, the instances of the document corresponding to respective different points in time;
generating a second language model for the document, the second language model comprising terms that occur in the instances of the document at or below the threshold frequency, terms in the first language model and terms in the second language model being non-overlapping; and
assigning the relevance weight to the document based upon the term being in the first language model rather than the second language model.

10. The method of claim 9 , further comprising:
assigning a first weight to the first language model based upon the query being classified as navigational;
assigning a second weight to the second language model based upon the query being classified as navigational; and
assigning the relevance weight to the document based upon the first weight assigned to the first language model and the second weight assigned to the second language model.

11. The method of claim 1 , wherein the term classification is additionally indicative of when the term was first introduced into the document in the window of time.

12. A method performed by at least one computing device that executes a search engine, the method comprising:
receiving, via an input interface of the search engine, a query that comprises a term;
generating a first language model for a document, the first language model comprises terms that occur in instances of the document above a threshold frequency, the instances of the document corresponding to respective different points in time;
generating a second language model for the document, the second language model comprising terms that occur in the instances of the document at or below the threshold frequency, terms in the first language model and terms in the second language model being non-overlapping; and
classifying the query as being navigational in nature;
assigning a first weight to the first language model based upon the query being classified as navigational;
assigning a second weight to the second language model based upon the query being classified as navigational;
assigning a relevance weight to the document based upon:
a term classification assigned to the term in the document, the term classification being representative of a length of time that the term has been present in the document over a window of time;
the term being in the first language model rather than the second language model; and
the first weight assigned to the first language model and the second weight assigned to the second language model, wherein the relevance weight is indicative of relevance of the document with respect to the query; and

causing a ranked list of search results to be presented on a display responsive to assigning the relevance weight to the document, the ranked list of search results comprises the document, wherein a position of the document in the ranked list of search results is based upon the relevance weight assigned to the document.

13. A system comprising:
at least one processor; and
memory that comprises instructions that, when executed by the at least one processor, cause the at least one processor to perform acts comprising:
receiving a query, the query comprising a term, wherein a search engine is configured to search over documents based upon the query;
assigning a classification to the query that indicates that the query is navigational;
assigning a relevance weight to a document in the documents, the relevance weight based upon:
the term being included in both the query and the document;
a classification assigned to the term in the document, the classification indicative of a length of time that the term has existed in the document over a window of time; and
the classification assigned to the query; and

causing a ranked list of search results to be presented on a display, wherein the search engine outputs the ranked list of search results based upon the query, and further wherein the document is positioned in the ranked list of search results based upon the relevance weight assigned to the document.

14. The system of claim 13 , wherein the acts are performed by a web-based search engine.

15. The system of claim 13 , wherein the acts are performed by a client application configured for execution on a personal computing device or an enterprise search tool.

16. The system of claim 13 , the acts further comprising assigning the relevance weight to the document based upon r

Assignees

• Dumais Susan T
• Elsas Jonathan Louis
• Liebling Daniel John
• Microsoft Technology Licensing Llc

Inventors

• Dumais Susan T
• Elsas Jonathan Louis
• Liebling Daniel John

Classifications

• G06F16/9535 Primary

  Search customisation based on user profiles and personalisation · CPC title

• G06F16/951 Primary

  Indexing; Web crawling techniques · CPC title