METHODS AND APPARATUS FOR CHARACTERIZING A SEARCH RESULT AS POTENTIAL SPAM

• US 20070185863A1
• Filed: 01/31/2007
• Published: 08/09/2007
• Est. Priority Date: 01/31/2006
• Status: Active Grant

• Current Assignee: Northwestern University, Taboola, Inc.
• Original Assignee: Media River, Inc.
• Inventors: Hammond, Kristian, Birnbaum, Lawrence, Budzik, Jerome, Crossen, Andrew, Sood, Sanjay
• Granted Patent: US 7,617,199 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/242 Query formulation
• CPC Class Codes: G06F 16/248 Presentation of query results
• CPC Class Codes: G06F 16/951 Indexing; Web crawling techniques
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/9538 Presentation of query results
• 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
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

Methods and apparatus assessing, ranking, organizing, and presenting search results associated with a user'"'"'s current work context are disclosed. The system disclosed assesses, ranks, organizes and presents search results against a user'"'"'s current work context by comparing statistical and heuristic models of the search results to a statistical and heuristic model of the user'"'"'s current work context. In this manner, search results are assessed, ranked, organized, and/or presented with the benefit of attributes of the user'"'"'s current work context that are predictive of relevance, such as words in a user'"'"'s document (e.g., web page or word processing document) that may not have been included in the search query. In addition, search results from multiple search engines are combined into an organization scheme that best reflects the user'"'"'s current task. As a result, lists of search results from different search engines can be more usefully presented to the user.

Claims

1. A method of characterizing a search result as potential spam, the method comprising:

receiving a search result from a search engine, the search result being based on a search query;

comparing data indicative of the search result to data indicative of the search query to determine a spam score, the spam score being based on at least one of a character length and a word length of a longest matching substring, the longest matching substring appearing in the search result and the search query;

comparing the spam score to a threshold; and

characterizing the search result as potential spam if the spam score crosses the threshold.

2. The method of claim 1, wherein the threshold is a tunable threshold.

3. The method of claim 1, wherein the spam score identifies a potential spam level.

4. The method of claim 1, wherein the search query is based on a first aspect of a user context, the first aspect of the user context including data indicative of text being accessed by a user, the query being different than the user context.

5. The method of claim 4, wherein the first aspect of the user context includes data indicative of the at least one task.

6. The method of claim 4, including comparing data indicative of a plurality of search results to data indicative of a second aspect of the user context to determine a plurality of relevance scores associated with the plurality of search results, the second aspect of the user context including data indicative of at least one task in which the user is engaged out of a plurality of possible user tasks.

7. The method of claim 6, wherein the second aspect of the user context includes the data indicative of the text being accessed by the user.

8. The method of claim 6, wherein the first aspect of the user context includes the second aspect of the user context.

9. The method of claim 6, wherein the first aspect of the user context is based on at least one of (a) a number of words in the text being accessed by the user, and (b) a number of sentences in the text being accessed by the user.

10. The method of claim 6, wherein the second aspect of the user context is based on at least one of (a) a location of the at least one predetermined word in the text being accessed by the user, (b) a style of the at least one predetermined word in the text being accessed by the user, (c) a presence of at least one specified word in the text being accessed by the user, (d) an absence of the at least one specified word in the text being accessed by the user, (e) metadata attributes of at least a portion of the text being accessed by the user, (f) a field presented by a computer application, (g) an attribute of information being presented in the computer application, (h) an element of the computer application visible to the user, (i) a document genre, (j) a document type, (k) a type associated with the computer application, (l) a method by which the user is accessing the computer application, (m) a role in an organization, (n) a type of the organization, (o) a property of the organization, (p) a stage in a task, (q) a stage in a workflow, (r) a type of task being supported by the computer application, (s) a stage in a task being executed by the computer application, (t) a pervious user behavior, (u) a topical area of interest, (v) a proportion of hyperlinked text to non-hyperlinked text, and (w) an average sentence length in the text being accessed by the user.

11. The method of claim 6, wherein the second aspect of the user context is based on at least five of (a) a location of the at least one predetermined word in the text being accessed by the user, (b) a style of the at least one predetermined word in the text being accessed by the user, (c) a presence of at least one specified word in the text being accessed by the user, (d) an absence of the at least one specified word in the text being accessed by the user, (e) metadata attributes of at least a portion of the text being accessed by the user, (f) a field presented by a computer application, (g) an attribute of information being presented in the computer application, (h) an element of the computer application visible to the user, (i) a document genre, (j) a document type, (k) a type associated with the computer application, (l) a method by which the user is accessing the computer application, (m) a role in an organization, (n) a type of the organization, (o) a property of the organization, (p) a stage in a task, (q) a stage in a workflow, (r) a type of task being supported by the computer application, (s) a stage in a task being executed by the computer application, (t) a pervious user behavior, (u) a topical area of interest, (v) a proportion of hyperlinked text to non-hyperlinked text, and (w) an average sentence length in the text being accessed by the user.

12. The method of claim 6, wherein the second aspect of the user context is based on (a) a style of the at least one predetermined word in the text being accessed by the user and (b) a type associated with a computer application.

13. The method of claim 6, including comparing data indicative of a plurality of search results to data indicative of at least one of the first aspect of the user context and the second aspect of the user context to determine a plurality of organization schemes, the plurality of organization schemes grouping at least a portion of the plurality of search results into at least two genres.

14. The method of claim 13, including generating a display indicative of a ranked list of search results that is indicative of the plurality of organization schemes.

15. The method of claim 1, including receiving a plurality of result models from the search engine, the plurality of result models including a plurality of terms associated with a plurality of search results and a plurality weights associated with the plurality of terms.

16. The method of claim 15, including comparing data indicative of the plurality of result models to data indicative of a user context to determine a plurality of scores associated with the plurality of search results.

17. The method of claim 1, including determining a second search query from a user context and interim search result.

18. An apparatus for characterizing a search result as potential spam, the apparatus comprising:

a processor;

a memory device operatively coupled to the processor; and

a network device operatively coupled to the processor;

wherein the memory device stores a software program to cause the processor to;

receive a search result from a search engine via the network device, the search result being based on a search query;

compare data indicative of the search result to data indicative of the search query to determine a spam score, the spam score being based on at least one of a character length and a word length of a longest matching substring, the longest matching substring appearing in the search result and the search query;

compare the spam score to a threshold; and

characterize the search result as potential spam if the spam score crosses the threshold.

19. The apparatus of claim 18, wherein the threshold is a tunable threshold.

20. The apparatus of claim 18, wherein the spam score identifies a potential spam level.

21. The apparatus of claim 18, wherein the search query is based on a first aspect of a user context, the first aspect of the user context including data indicative of text being accessed by a user, the query being different than the user context.

22. The apparatus of claim 21, wherein the first aspect of the user context includes data indicative of the at least one task.

23. The apparatus of claim 21, wherein the software program is structured to cause the processor to compare data indicative of a plurality of search results to data indicative of a second aspect of the user context to determine a plurality of relevance scores associated with the plurality of search results, the second aspect of the user context including data indicative of at least one task in which the user is engaged out of a plurality of possible user tasks.

24. The method of claim 23, wherein the second aspect of the user context is based on at least five of (a) a location of the at least one predetermined word in the text being accessed by the user, (b) a style of the at least one predetermined word in the text being accessed by the user, (c) a presence of at least one specified word in the text being accessed by the user, (d) an absence of the at least one specified word in the text being accessed by the user, (e) metadata attributes of at least a portion of the text being accessed by the user, (f) a field presented by a computer application, (g) an attribute of information being presented in the computer application, (h) an element of the computer application visible to the user, (i) a document genre, (j) a document type, (k) a type associated with the computer application, (l) a method by which the user is accessing the computer application, (m) a role in an organization, (n) a type of the organization, (o) a property of the organization, (p) a stage in a task, (q) a stage in a workflow, (r) a type of task being supported by the computer application, (s) a stage in a task being executed by the computer application, (t) a pervious user behavior, (u) a topical area of interest, (v) a proportion of hyperlinked text to non-hyperlinked text, and (w) an average sentence length in the text being accessed by the user.

25. The method of claim 23, wherein the second aspect of the user context is based on (a) a style of the at least one predetermined word in the text being accessed by the user and (b) a type associated with a computer application.

31. The computer readable medium of claim 21, wherein the software program is structured to cause the computing device to compare data indicative of a plurality of search results to data indicative of a second aspect of the user context to determine a plurality of relevance scores associated with the plurality of search results, the second aspect of the user context including data indicative of at least one task in which the user is engaged out of a plurality of possible user tasks.

32. The method of claim 31, wherein the second aspect of the user context is based on at least five of (a) a location of the at least one predetermined word in the text being accessed by the user, (b) a style of the at least one predetermined word in the text being accessed by the user, (c) a presence of at least one specified word in the text being accessed by the user, (d) an absence of the at least one specified word in the text being accessed by the user, (e) metadata attributes of at least a portion of the text being accessed by the user, (f) a field presented by a computer application, (g) an attribute of information being presented in the computer application, (h) an element of the computer application visible to the user, (i) a document genre, (j) a document type, (k) a type associated with the computer application, (l) a method by which the user is accessing the computer application, (m) a role in an organization, (n) a type of the organization, (o) a property of the organization, (p) a stage in a task, (q) a stage in a workflow, (r) a type of task being supported by the computer application, (s) a stage in a task being executed by the computer application, (t) a pervious user behavior, (u) a topical area of interest, (v) a proportion of hyperlinked text to non-hyperlinked text, and (w) an average sentence length in the text being accessed by the user.

33. The method of claim 31, wherein the second aspect of the user context is based on (a) a style of the at least one predetermined word in the text being accessed by the user and (b) a type associated with a computer application.

26. A computer readable medium storing a software program to cause a computing device to:

receive a search result from a search engine via the network device, the search result being based on a search query;

compare data indicative of the search result to data indicative of the search query to determine a spam score, the spam score being based on at least one of a character length and a word length of a longest matching substring, the longest matching substring appearing in the search result and the search query;

compare the spam score to a threshold; and

characterize the search result as potential spam if the spam score crosses the threshold.

27. The computer readable medium of claim 26, wherein the threshold is a tunable threshold.

28. The computer readable medium of claim 26, wherein the spam score identifies a potential spam level.

29. The computer readable medium of claim 26, wherein the search query is based on a first aspect of a user context, the first aspect of the user context including data indicative of text being accessed by a user, the query being different than the user context.

30. The computer readable medium of claim 29, wherein the first aspect of the user context includes data indicative of the at least one task.