AUTOMATIC CUSTOMIZATION AND RENDERING OF ADS BASED ON DETECTED FEATURES IN A WEB PAGE

• US 20090248484A1
• Filed: 03/28/2008
• Published: 10/01/2009
• Est. Priority Date: 03/28/2008
• Status: Abandoned Application

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Surendran, Arungunram C., Zhang, Yi, Dominowska, Ewa
• Application Number:
  US12/057,871
• Publication Number:
  US 20090248484A1
• Time in Patent Office: Days
• US Class Current: 705/10
• CPC Class Codes: G06Q 30/0201 Market modelling; Market analysis; Collecting market data
• CPC Class Codes: G06Q 30/0255 based on user history
• CPC Class Codes: G06Q 30/0263 based upon Internet or website rating

Abstract

Media and system are provided for customizing ads for delivery based on features in a web page. Text in a web page is evaluated to determine the sentiment of the web page. The sentiment can be based on the overall web page or on individual blocks of text within the web page. Based on the sentiment, an ad is created from the text in the web page. The resulting ad is rendered in the web page and shown to a user.

Claims

1. One or more computer-readable storage media having computer instructions embodied thereon for performing a method for tailoring advertisements to a web page at a time of delivery, the method comprising:

evaluating one or more words in the web page to determine separately a product for advertising;

determining one or more sentiments about the product from an evaluation of the one or more words;

creating at least one of an extracted summary and a descriptive summary based on the one or more sentiments;

placing the at least one of the extracted summary and the descriptive summary into a template to form an advertisement focused on the product; and

rendering the advertisement in the web page.

2. The media of claim 1, wherein the extracted summary is a phrase taken from information shown in the web page.

3. The media of claim 2, wherein the descriptive summary is a set of phrases created from a variety of text shown in the web page.

4. The media of claim 1, wherein determining the one or more sentiments comprises determining whether one or more persons like or dislike the product.

5. The media of claim 4, wherein determining the one or more sentiments comprises operating a natural language processing algorithm to determine a sentiment from the one or more words.

6. The media of claim 4, further comprising rendering the advertisement in the web page near a set of text with a sentiment where a person likes the product.

7. The media of claim 4, further comprising associating the one or more sentiments to one or more objects of the product wherein the one or more objects are aspects of the product or features of the product.

8. One or more computer-readable storage media having computer instructions embodied thereon for performing a method for determining how an ad is to appear in a web page, the method comprising:

evaluating, in the web page, text that discusses a product;

grouping the text into one or more blocks wherein the one or more blocks discuss the product;

classifying each of the one or more blocks as having a positive sentiment, a negative sentiment, or a neutral sentiment for the product or an aspect of the product;

determining whether the ad is to be placed in the web page based on one or more sentiments assigned to the one or more blocks;

if the ad is to be placed in the web page, creating the ad with one or more words from the text of the web page; and

rendering the ad in the web page.

9. The media of claim 8, further comprising deciding to place the ad in the web page when the one or more blocks have a majority of the positive sentiment.

10. The media of claim 9, further comprising deciding not to place the ad in the web page when the one or more blocks have a majority of the negative sentiment.

11. The media of claim 10, further comprising deciding to place a competitor ad rather than the ad in the web page.

12. The media of claim 8, wherein creating the ad with the one or more words comprises filling in an ad template with the one or more words.

13. The media of claim 12, wherein the ad template is provided by advertisers.

14. The media of claim 8, further comprising selecting one ad template from one or more ad templates provided by an advertiser.

15. The media of claim 14, further comprising customizing the ad template for the ad to be displayed in the web page.

16. The media of claim 8, further comprising locating the ad near a member of the one or more blocks that have the positive sentiment.

17. A system having a computer with a processor and a memory, to determine an ad for display in a web page, comprising:

determining a set of advertisers that desire to place ads in the web page wherein the advertisers provide a name of a product, a link to a product web page of the product, and one or more keywords associated with the product;

determining a first subset of the set of advertisers based on at least one of an amount paid by an advertiser to display the ad, a probability of the ad being selected by a user, and a relevance of the one or more keywords to the web page;

determining a second subset of the set of advertisers from the first subset based on an evaluation of the web page wherein one or more sentiments about one or more products or an aspect of the one or more products are determined from the evaluation;

selecting the advertiser wherein the one or more sentiments are positive about the product or the aspect of the product of the advertiser;

creating the ad for the advertiser with one or more words in the web page;

and rendering the ad in the web page.

18. The system of claim 17, wherein creating the ad comprises filling in an ad template with information about the product.

19. The system of claim 17, further comprising locating the ad near a set of text in the web page wherein the set of text includes a sentiment that is positive.

20. The system of claim 17, wherein the evaluation of the web page comprises an analysis of the one or more words in the web page.