Search engine spam detection using external data

• US 20060004748A1
• Filed: 05/21/2004
• Published: 01/05/2006
• Est. Priority Date: 05/21/2004
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Watson, Eric B., Ramarathnam, Bama, Crumb, Janine Ruth
• Granted Patent: US 7,349,901 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/9535 Search customisation based on user profiles and personalisation
• CPC Class Codes: G06F 16/9536 Search customisation based on social or collaborative filtering
• CPC Class Codes: G06F 16/9538 Presentation of query results
• CPC Class Codes: Y10S 707/913 Multimedia
• CPC Class Codes: Y10S 707/959 Network
• CPC Class Codes: Y10S 707/99935 Query augmenting and refining, e.g. inexact access
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

Evaluating an electronic document in connection with a search. An external source provides data for use in evaluating an electronic document retrieved by a search engine. A first confidence level of the electronic document is determined based on the externally provided data. The first confidence level indicates a likelihood that the electronic document is undesirable. A second confidence level of the electronic document is determined based on attributes of the electronic document. The second confidence level indicates a likelihood that the electronic document is unsatisfactory with respect to a search. A rating for the electronic document generated as a function of the determined first confidence level and the determined second confidence level is used to categorize the electronic document as unsatisfactory in connection with a received search request.

Claims

1. A method of evaluating an electronic document in connection with a search, said method comprising:

determining a first confidence level of an electronic document, said electronic document being retrievable by a search engine in response to a search request from a user, said first confidence level indicating a likelihood that the electronic document is undesirable based on information provided by a source external to the search engine;

determining a second confidence level of the electronic document, said second confidence level indicating a likelihood that the electronic document is unsatisfactory with respect to the search request based on one or more attributes of the electronic document;

generating a rating for the electronic document as a function of the determined first confidence level and the determined second confidence level; and

designating the electronic document as unsatisfactory in connection with the search request based on the generated rating of the electronic document.

2. The method of claim 1, wherein said external source comprises an electronic mail spam detection system.

3. The method of claim 1, wherein said electronic document comprises one or more of the following:

a web page and a multimedia file.

4. The method of claim 1, wherein determining the first confidence level comprises:

receiving one or more host names from the external source, wherein information provided by the received host names has a predetermined likelihood of being undesirable;

identifying the electronic document as being provided by one of the received host names; and

specifying the first confidence level for the electronic document in response to identifying the electronic document as being provided by one of the received host names, said first confidence level based on the predetermined likelihood.

5. The method of claim 4, further comprising specifying the first confidence level for one or more other electronic documents linked from the electronic document.

6. The method of claim 1, wherein determining the first confidence level comprises:

receiving one or more network addresses from the external source, wherein the external source identifies one or more electronic documents located at one of the received network addresses as having a predetermined likelihood of being undesirable;

identifying the electronic document as being located at one of the received network addresses; and

specifying the first confidence level for the electronic document in response to identifying the electronic document as being located at one of the received network addresses, said first confidence level based on the predetermined likelihood.

7. The method of claim 1, wherein determining the first confidence level comprises:

receiving user-provided information from the external source, said user-provided information specifying the electronic document as undesirable;

identifying one or more attributes of the electronic document in response to the received user-provided information, said attributes characterizing an undesirable pattern of the electronic document; and

specifying the first confidence level for the electronic document based on the identified attributes of the electronic document.

8. The method of claim 1, wherein determining the first confidence level comprises:

receiving one or more terms from the external source, wherein the external source determines that one or more electronic documents in which at least one of the received terms appears have a predetermined likelihood of being undesirable;

detecting an appearance of at least one of the received terms in the electronic document; and

specifying the first confidence level for the electronic document in response to detecting the appearance of at least one of the received terms in the electronic document, said first confidence level based on the predetermined likelihood.

9. The method of claim 1, wherein determining the second confidence level comprises:

parsing the electronic document to identify the attributes of the electronic document, said attributes characterizing an undesirable pattern of the electronic document with respect to a search; and

determining the second confidence level based on the identified attributes of the electronic document.

10. The method of claim 1, wherein determining the second confidence level comprises:

receiving user-provided information with respect to the electronic document, said received user-provided information specifying the electronic document as undesirable in a search result;

identifying one or more attributes of the electronic document in response to the received user-provided information, said attributes characterizing an undesirable pattern of the electronic document with respect to a search; and

determining the second confidence level based on the identified attributes of the electronic document.

11. The method of claim 1, further comprising:

providing a search result to the user in response to the received search request; and

performing one or more of the following;

demoting the electronic document designated as unsatisfactory in the provided search result, excluding the electronic document designated as unsatisfactory from the provided search result, and preserving a ranking of the electronic document in the provided search result when the ranking of the electronic document exceeds a predetermined rank in the provided search result.

12. The method of claim 1, wherein one or more computer-readable media have computer-executable instructions for performing the method recited in claim 1.

13. A method of evaluating an electronic document in connection with a search, said method comprising:

receiving user-provided information with respect to an electronic document, said electronic document being retrievable by a search engine in response to a search request from a user, said user-provided information characterizing the electronic document as undesirable;

generating a rating for the electronic document as a function of the received user-provided information; and

designating the electronic document as unsatisfactory in connection with the search request according to the generated rating of the electronic document.

14. The method of claim 13, wherein said electronic document comprises one or more of the following:

a web page and a multimedia file.

15. The method of claim 13, wherein the received user-provided information specifies the electronic document as being associated with an undesirable electronic mail.

16. The method of claim 13, wherein the received user-provided information specifies the electronic document as being undesirable in a search result.

17. The method of claim 13, wherein generating the rating for the electronic document comprises:

parsing the electronic document to identify one or more attributes of the electronic document; and

applying the identified attributes to a probabilistic classifier to generate the rating for the electronic document, said probabilistic classifier being trained to recognize if the identified attributes are undesirable.

18. The method of claim 13, wherein the probabilistic classifier comprises one or more classifiers selected from a group comprising:

a Naï

ve Bayesian classifier, a limited dependence Bayesian classifier, a Bayesian network classifier, a decision tree, a support vector machine, a content matching classifier, a maximum entropy classifier, and a combination thereof.

19. The method of claim 13, wherein generating the rating for the electronic document comprises:

determining trustworthiness of the received user-provided information, and

generating the rating for the electronic document as a function of the determined trustworthiness.

20. The method of claim 19, wherein determining the trustworthiness of the received user-provided information comprises:

receiving other user-provided information with respect to the electronic document, and

determining if the other user-provided information corresponds with the received user-provided information.

21. The method of claim 13, wherein one or more computer-readable media have computer-executable instructions for performing the method recited in claim 13.

22. A system for evaluating an electronic document in connection with a search, said system comprising:

a processor for receiving a search request from a user and for identifying an electronic document based on the received search request;

a memory area storing data provided by a source external to the processor for use in evaluating whether the electronic document is undesirable;

said processor being configured to determine a first confidence level of the electronic document, said first confidence level indicating a likelihood that the electronic document is undesirable based on the data provided by the external source;

said processor being further configured to establish a second confidence level of the electronic document, said second confidence level indicating a likelihood that the electronic document is unsatisfactory with respect to a search based on one or more attributes of the electronic document;

said processor being further configured to generate a rating for the electronic document as a function of the determined first confidence level and the established second confidence level and to categorize the electronic document as unsatisfactory in connection with the received search request based on the generated rating of the electronic document.

23. The system of claim 22, wherein said external source comprises an electronic mail spam detection system.

24. The system of claim 22, wherein the data provided by the external source identifies one or more host names, each of said host names providing information having a predetermined likelihood of being undesirable, wherein the processor is configured to identify the electronic document as being provided by one of the host names, and wherein the processor is further configured to specify the first confidence level for the electronic document in response to identifying the electronic document as being provided by one of the host names, said first confidence level being based on the predetermined likelihood.

25. The system of claim 22, wherein the data provided by the external source identifies one or more network addresses, wherein the external source identifies one or more electronic documents located at one of the network addresses as having a predetermined likelihood of being undesirable, wherein the processor is configured to identify the electronic document as being located at one of the network addresses, and wherein the processor is configured to specify the first confidence level for the electronic document in response to identifying the electronic document as being located at one of the network addresses, said first confidence level being based on the predetermined likelihood.

26. The system of claim 22, wherein the data provided by the external source identifies one or more terms, wherein the external source determines that one or more electronic documents in which at least one of the terms appears have a predetermined likelihood of being undesirable, wherein the processor is configured to determine when at least one of the terms appears in the electronic document, and wherein the processor is configured to specify the first confidence level for the electronic document in response to determining that at least one of the terms appears in the electronic document, said first confidence level being based on the predetermined likelihood.

27. The system of claim 22, wherein the processor is configured to parse the electronic document to identify the attributes of the electronic document and to establish the second confidence level based on the identified attributes of the electronic document, said identified attributes characterizing an undesirable pattern of the electronic document with respect to the search.

28. The system of claim 22, wherein the processor is configured to provide a search result to the user in response to the received search request and to perform one or more of the following:

demoting the electronic document categorized as unsatisfactory in the provided search result, excluding the electronic document categorized as unsatisfactory from the provided search result, and preserving a ranking of the electronic document in the provided search result when the ranking of the electronic document exceeds a predetermined rank in the provided search result.

29. One or more computer-readable media having computer-executable components for evaluating an electronic document in connection with a search, said computer-readable media comprising:

an interface component for receiving user-provided information with respect to an electronic document, said electronic document being retrievable in response to a search request from a user, said user-provided information characterizing the electronic document as undesirable;

an analyzing component for generating a rating for the electronic document as a function of the received user-provided information; and

a query component for categorizing the electronic document as unsatisfactory in connection with the search request according to the generated rating of the electronic document.

30. The computer-readable media of claim 29, wherein the received user-provided information specifies that the electronic document is associated with a source of an undesirable electronic mail.

31. The computer-readable media of claim 29, wherein the received user-provided information specifies that the electronic document is undesirable in a search result.

32. The computer-readable media of claim 29, wherein the analyzing component is configured to parse the electronic document to identify one or more attributes of the electronic document, and wherein the analyzing component is further configured to apply the identified attributes to a probabilistic classifier to generate the rating for the electronic document, said probabilistic classifier being trained to recognize if the identified attributes are undesirable.

33. The computer-readable media of claim 32, wherein the probabilistic classifier comprises one or more classifiers selected from a group comprising:

a Naï

ve Bayesian classifier, a limited dependence Bayesian classifier, a Bayesian network classifier, a decision tree, a support vector machine, a content matching classifier, a maximum entropy classifier, and a combination thereof.

34. The computer-readable media of claim 29, wherein the analyzing component is configured to determine trustworthiness of the received user-provided information and to generate the rating for the electronic document as a function of the determined trustworthiness.

35. The computer-readable media of claim 34, wherein the interface component is configured to receive other user-provided information with respect to the electronic document, and wherein the analyzing component is configured to examine if the other user-provided information corresponds to the received user-provided information to determine the trustworthiness of the received electronic document.

36. One or more computer-readable media having computer-executable components for evaluating an electronic document in connection with a search, said computer-readable media comprising:

a query component to receive a search request from a user and to identify an electronic document based on the received search request;

an external component to provide data for use in evaluating whether the electronic document is undesirable;

an internal component to determine a first confidence level of the electronic document, said first confidence level indicating a likelihood that the electronic document is undesirable based on the data provided by the external component, said internal component being further configured to establish a second confidence level of the electronic document, said second confidence level indicating a likelihood that the electronic document is unsatisfactory with respect to a search based on one or more attributes of the electronic document;

an analyzing component to generate a rating for the electronic document as a function of the determined first confidence level and the established second confidence level; and

wherein the query component is configured to classify the electronic document as unsatisfactory in connection with the received search request based on the generated rating of the electronic document.

37. The computer-readable media of claim 36, wherein the data provided by the external component identifies one or more host names, each of said host names providing information having a predetermined likelihood of being undesirable, wherein the internal component is configured to identify the electronic document as being provided by one of the host names, and wherein the internal component is further configured to specify the first confidence level for the electronic document in response to identifying the electronic document as being provided by one of the host names, said first confidence level being based on the predetermined likelihood.

38. The computer-readable media of claim 36, wherein the data provided by the external component identifies one or more network addresses, wherein the external component identifies one or more electronic documents located at one of the network addresses as having a predetermined likelihood of being undesirable, wherein the internal component is configured to identify the electronic document as being located at one of the network addresses, and wherein the internal component is configured to specify the first confidence level for the electronic document in response to identifying the electronic document as being located at one of the network addresses, said first confidence level being based on the predetermined likelihood.

39. The computer-readable media of claim 36, wherein the data provided by the external component identifies one or more terms, wherein the external component determines that one or more electronic documents in which at least one of the terms appears have a predetermined likelihood of being undesirable, wherein the internal component is configured to determine when at least one of the terms appears in the electronic document, and wherein the internal component is configured to specify the first confidence level for the electronic document in response to determining that at least one of the terms appears in the electronic document, said first confidence level being based on the predetermined likelihood.

40. The computer-readable media of claim 36, wherein the query component is configured to provide a search result to the user in response to the received search request and to perform one or more of the following:

demoting the electronic document classified as unsatisfactory in the provided search result, excluding the electronic document classified as unsatisfactory from the provided search result, and preserving a ranking of the electronic document in the provided search result when the ranking of the electronic document exceeds a predetermined rank in the provided search result.