US 7,594,013 B2
Creating home pages based on user-selected information of web pages
Jian Wang, Beijing (China); Hua-Jun Zeng, Beijing (China); Chenxi Lin, Beijing (China); Zheng Chen, Beijing (China); Benyu Zhang, Beijing (China); and Bing Sun, Beijing (China)
Assigned to Microsoft Corporation, Redmond, Wash. (US)
Filed on May 24, 2005, as Appl. No. 11/136,029.
Prior Publication US 2006/0271834 A1, Nov. 30, 2006
Int. Cl. G06F 15/173 (2006.01)
U.S. Cl. 709—224  [709/223; 709/225; 709/226] 8 Claims
OG exemplary drawing
 
1. A method performed by a computer of creating a home page, the method comprising the steps of:
for each of a plurality of web pages,
receiving from a user a selection of the web page, the selected web page having an original tree of nodes;
displaying to the user the selected web page;
receiving from the user a selection of information of the displayed web page, the selected information to be located on the home page;
defining a web block of the displayed web page based on the selected information of the displayed web page, the web block representing an original subtree of the original tree structure of the displayed web page; and
adding the defined web block to the home page;
when displaying the home page, displaying the defined web block on the home page for each of the plurality of selected web pages; and
for each of the selected web pages,
tracking by the computer the selected information of the selected web page, the tracking including:
identifying a mapping of nodes of the original tree of the selected web page to nodes of a new tree of the selected web page that minimizes differences between the original subtree of the original tree and a corresponding subtree of the new tree, the original tree and the original subtree and the new tree and the corresponding subtree are based on the document object model (DOM); and
collecting user feedback relating to structural features and content features of the original subtree and the corresponding subtree, the structural features including at least one feature selected from a group consisting of root level, RankInSibling, and Offspringnumbers, the content features including at least one feature selected from the group consisting of Text Length, Key words, Title, LineNum, and ImageInfo;
learning via a learning algorithm from the collected user feedback features that indicate which information of the new tree correspond to the selected information of the original tree; and
identifying information of the selected web page that corresponds to the previously selected information of the selected web page based on the learning and the mapping; and
updating the web blocks added to the home page based on the tracking of the selected information of the selected web page.