Annotating HTML segments with functional labels

• US 9,594,730 B2
• Filed: 07/01/2010
• Issued: 03/14/2017
• Est. Priority Date: 07/01/2010
• Status: Expired due to Fees

• Current Assignee: R2 Solutions LLC (Acacia Research Corporation)
• Original Assignee: Yahoo! Inc. (Apollo Global Management, Inc.)
• Inventors: Rajan, Suju, Gaffney, Scott J., Punera, Kunal
• Primary Examiner(s): STORK, KYLE R
• Application Number:
  US12/829,265
• Publication Number:
  US 20120005686A1
• Time in Patent Office: 2,448 Days
• US Class Current: 1/1
• CPC Class Codes: G06F 40/137 Hierarchical processing, e.g. outlines

Abstract

A method and apparatus is described for assigning functional labels to segments of web pages in an application-independent way. In the approach described herein, one of a generic set functional labels are automatically assigned to each segment of a web page, where the generic functional labels may be topic-independent and application-independent. Applications with different needs can determine which segments of the web page to process based on which functional labels correspond to the types of information needed by each application. Thus, the work of classifying the function of each segment of a web page is separated from the work of selecting which segments satisfy the need of a particular application. The work of classification can be performed in an application-independent way, relieving the burden from every application developer from having to create their own classifiers.

Claims

1. A method comprising:

processing a web page to determine a plurality of segments, wherein each segment from the plurality of segments includes one or more HTML elements;

each machine-based classifier of a plurality of machine-based classifiers generating, based at least upon metadata associated with two or more segments from the plurality of segments that indicates one or more presentation features in the HTML elements of the two or more segments from the plurality of segments, a probability output for each segment of the two or more segments from the plurality of segments, wherein each functional category from the plurality of functional categories corresponds to a functional role of HTML elements in the web page;

wherein each machine-based classifier from the plurality of machine-based classifiers corresponds to a functional category from the plurality of functional categories;

assigning, based on the plurality of probability output, one or more functional categories to each segment of the two or more segments;

a first application selecting a first set of functional categories from the plurality of functional categories;

a second application that is different than the first application selecting a second set of functional categories from the plurality of functional categories, wherein the second set of functional categories does not include functional categories from the first set of functional categories;

the first application selecting for processing, based upon the first set of functional categories and the functional categories assigned to the two or more segments, a first set of one or more segments from the two or more segments;

the second application selecting for processing, based upon the second set of functional categories and the functional categories assigned to the two or more segments, a second set of one or more segments from the two or more segments, wherein the second set of one or more segments includes at least one segment that is not in the first set of one or more segments and the first set of one or more segments includes at least one segment that is not in the second set of one or more segments;

the first application processing content contained in the first set of one or more segments and not processing content contained in the second set of one or more segments;

the second application processing content contained in the second set of one or more segments and not processing content contained in the first set of one or more segments; and

wherein the method is performed by one or more computing devices.

2. The method of claim 1, wherein the first set of one or more segments does not include a segment assigned a main content functional category, wherein segments assigned the main content functional category contain the most text that is relevant to the topic of the web page.

3. The method of claim 1, wherein the second application indexes the web page for a search engine by ignoring the content of each segment from the two or more segments that is assigned an advertisement functional category or a site navigation functional category.

4. The method of claim 1, wherein the functional categories assigned to the two or more segments are independent of a topic of one or more topics associated with the web page.

5. The method of claim 1, wherein the plurality of functional categories includes at least one of:

a) user-generated content;

b) site navigation;

or

c) boiler-plate.

6. The method of claim 1, wherein the first application determines a topic for the web page by processing the content contained in the first set of one or more segments, wherein the first set of one or more segments includes segments that are assigned a functional category of:

a) main content

b) content pointers; and

c) site navigation;

wherein the first set of one or more segments does not include segments that are assigned a functional category of

a) advertisement;

or

b) user-generated content.

7. The method of claim 1, wherein the first application determines a topic for the web page by analyzing a frequency of words that appear as text within the first set of one or more segments;

and

wherein the frequency of words that appear as text within segments of the first set of one or more segments assigned an advertisement functional category or user-generated content functional category are not analyzed.

8. The method of claim 1, wherein the metadata associated with the two or more segments includes at least one of:

a) amount of screen space taken to display the segment;

b) height or width of the screen space taken to display the segment;

c) number of hyperlinks contained within the segment;

d) amount of text within the segment;

e) ratio of hyperlinks to text within the segment;

f) types of html elements contained within the segment;

or

g) color of the section and whether the color of the segment is different from the color of other segments.

9. The method of claim 1, wherein assigning one or more functional categories further comprises:

training a machine-based classifier based on the metadata associated with the two or more segments and functional category data provided by human editors;

the machine-based classifier improving the accuracy of functional category assignment based on regression techniques.

10. The method of claim 1, wherein the first application is a web crawler and the first set of one or more segments comprises a set of segments that are assigned a main content functional category.

11. One or more non-transitory computer-readable media storing instructions which, when processed by one or more processors, cause:

processing a web page to determine a plurality of segments, wherein each segment from the plurality of segments includes one or more HTML elements;

each machine-based classifier of a plurality of machine-based classifiers generating, based at least upon metadata associated with two or more segments from the plurality of segments that indicates one or more presentation features in the HTML elements of the two or more segments from the plurality of segments, a probability output for each segment of the two or more segments from the plurality of segments, wherein each functional category from the plurality of functional categories corresponds to a functional role of HTML elements in the web page;

wherein each machine-based classifier from the plurality of machine-based classifiers corresponds to a functional category from the plurality of functional categories;

assigning, based on the plurality of probability output, one or more functional categories to each segment of the two or more segments;

a first application selecting a first set of functional categories from the plurality of functional categories;

a second application that is different than the first application selecting a second set of functional categories from the plurality of functional categories, wherein the second set of functional categories does not include functional categories from the first set of functional categories;

the first application selecting for processing, based upon the first set of functional categories and the functional categories assigned to the two or more segments, a first set of one or more segments from the two or more segments;

the second application selecting for processing, based upon the second set of functional categories and the functional categories assigned to the two or more segments, a second set of one or more segments from the two or more segments, wherein the second set of one or more segments includes at least one segment that is not in the first set of one or more segments and the first set of one or more segments includes at least one segment that is not in the second set of one or more segments;

the first application processing content contained in the first set of one or more segments and not processing content contained in the second set of one or more segments; and

the second application processing content contained in the second set of one or more segments and not processing content contained in the first set of one or more segments.

12. The one or more non-transitory computer-readable media of claim 11, wherein the first set of one or more segments does not include a segment assigned a main content functional category, wherein segments assigned the main content functional category contain the most text that is relevant to the topic of the web page.

13. The one or more non-transitory computer-readable media of claim 11, wherein the second application indexes the web page for a search engine by ignoring the content of each segment from the two or more segments that is assigned an advertisement functional category or a site navigation functional category.

14. The one or more non-transitory computer-readable media of claim 11, wherein the functional categories assigned to the two or more segments are independent of a topic of one or more topics associated with the web page.

15. The one or more non-transitory computer-readable media of claim 11, wherein the plurality of functional categories includes at least one of:

a) user-generated content;

b) site navigation;

or

c) boiler-plate.

16. The one or more non-transitory computer-readable media of claim 11, wherein the first application determines a topic for the web page by processing the content contained in the first set of one or more segments, wherein the first set of one or more segments includes segments that are assigned a functional category of:

a) main content

b) content pointers; and

c) site navigation;

wherein the first set of one or more segments does not include segments that are assigned a functional category of

a) advertisement;

or

b) user-generated content.

17. The one or more non-transitory computer-readable media of claim 11, wherein the first application determines a topic for the web page by analyzing a frequency of words that appear as text within the first set of one or more segments;

and

wherein the frequency of words that appear as text within segments of the first set of one or more segments assigned an advertisement functional category or user-generated content functional category are not analyzed.

18. The one or more non-transitory computer-readable media of claim 11, wherein the metadata associated with the two or more segments includes at least one of:

a) amount of screen space taken to display the segment;

b) height or width of the screen space taken to display the segment;

c) number of hyperlinks contained within the segment;

d) amount of text within the segment;

e) ratio of hyperlinks to text within the segment;

f) types of html elements contained within the segment;

or

g) color of the section and whether the color of the segment is different from the color of other segments.

19. The one or more non-transitory computer-readable media of claim 11, wherein assigning one or more functional categories further comprises:

training a machine-based classifier based on the metadata associated with the two or more segments and functional category data provided by human editors;

the machine-based classifier improving the accuracy of functional category assignment based on regression techniques.

20. The one or more non-transitory computer-readable media of claim 11, wherein the first application is a web crawler and the first set of one or more segments comprises a set of segments that are assigned a main content functional category.