System and methods for indentifying the potential advertising value of terms found on web pages

• US 7,685,197 B2
• Filed: 05/05/2005
• Issued: 03/23/2010
• Est. Priority Date: 05/05/2005
• Status: Expired due to Fees

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Pierce, Thomas, Fain, Daniel C.
• Primary Examiner(s): Vital; Pierre M
• Assistant Examiner(s): OBISESAN, AUGUSTINE KUNLE
• Application Number:
  US11/122,780
• Publication Number:
  US 20060253437A1
• Time in Patent Office: 1,783 Days
• Field of Search: 707/5, 707/2
• US Class Current: 707/732
• 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: G06Q 30/02 Marketing; Price estimation or determination; Fundraising
• CPC Class Codes: G06Q 30/0256 User search
• CPC Class Codes: G06Q 30/0257 User requested

Abstract

The present invention provides a method and system for identifying terms in a requested Web document which can be used to identify other web page documents in which the user is likely to have an interest. Terms on previously prepared lists are compared to the content of the Web page document, and where matches are found, mathematical techniques including linear regression can be used to determine the relevance of the textual term to the content of the page. This can be used to suggest other linked Web page documents in which the user is likely to have an interest. Where the prepared lists include a term that has been purchased by advertisers, and there is a match for that term on the Web page document, an ad which is likely to interest the reader can be displayed on the requested Web page document, and can include a link to the advertiser'"'"'s web site.

Claims

1. A computer-implemented method for estimating an interest a reader has in a requested document, useful in association with a list of textual terms with advertising value, the method comprising:

building a trained model comprising using linear regression to identify a scoring function based on textual terms and associated documents,

wherein the scoring function is of the form;

o(k,p)=α

+Σ

β

_i log f_i(k)+Σ

γ

_i log g_i(k,p)

wherein β

_i and γ

_i are comparative relevance attributes and α

is a relevancy factor;

wherein k is a term factor and p is a web document identifier,

eliminating extraneous material from the requested document;

comparing a plurality of textual terms of the requested document to the list of textual terms with the advertising value, wherein the advertising value is based upon a bid value associated with one or more textual terms of the list of textual terms by one or more advertisers and the textual terms also including non-contextual features;

computing a significance level based on the trained model for each of the plurality of textual terms of the requested document found on the list of textual terms with advertising value according to the advertising value of each of the plurality of terms of the requested document and computing the significant level based on the non-contextual feature of the positions of the textual terms in the requested document;

ranking each of the plurality of textual terms of the requested document found on the list of textual terms with advertising value according to the significance level of each of the plurality of textual terms of the requested document that are found on the list of textual terms with advertising value;

selecting one or more of the plurality of ranked textual terms as keywords to search for one or more relevant advertisements; and

displaying the one or more relevant advertisements along with the requested document.

2. The method of claim 1 wherein the plurality of textual terms of the requested document include individual alphabetic words.

3. The method of claim 1 wherein the plurality of textual terms of the requested document include groups of at least two contiguous alphabetic words.

4. The method of claim 1 wherein the plurality of textual terms of the requested document include groups of at least two non-contiguous alphabetic words.

5. The method of claim 1 wherein the plurality of textual terms of the requested document include fragments of alphabetic words.

6. The method of claim 1 wherein the plurality of textual terms of the requested document include numbers.

7. The method of claim 1 wherein the plurality of textual terms of the requested document include individual non-alphabetic words.

8. The method of claim 1 wherein the plurality of textual terms of the requested document include groups of at least two contiguous non-alphabetic words.

9. The method of claim 1 wherein the plurality of textual terms of the requested document include groups of at least two non-contiguous non-alphabetic words.

10. The method of claim 1 wherein the requested document is a document accessed from a wide area network.

11. The method of claim 1 wherein features of each of the plurality of textual terms of the requested document found on the list of textual terms used by the learning machine to compute the significance level for each of the textual terms of the requested document found on the list of textual terms include contextual features.

12. The method of claim 1 wherein ranking of each of the plurality of textual terms of the requested document found on the list of trained textual terms according to the significance level of each of the plurality of textual terms of the requested document found on the list of textual terms includes ranking by advertising value.