| US 7,590,628 B2 | ||
| Determining document subject by using title and anchor text of related documents | ||
| Shubin Zhao, Jersey City, N.J. (US) | ||
| Assigned to Google, Inc., Mountain View, Calif. (US) | ||
| Filed on Mar. 31, 2006, as Appl. No. 11/394,610. | ||
| Prior Publication US 2007/0240031 A1, Oct. 11, 2007 | ||
| Int. Cl. G06F 17/00 (2006.01) | ||
| U.S. Cl. 707—6 [715/205] | 14 Claims |
| 1. A computer-implemented method of determining the subject of a target document, comprising: using a computer processor to
perform:
(a) identifying a plurality of peer documents within the same domain as the target document, each of the plurality of peer
documents being associated with the target document and being stored in a computer memory;
(b) for each of the plurality of peer documents,
(i) identifying a plurality of linking documents having links that link to the peer document, each link having an anchor text,
(ii) identifying a plurality of linking documents having links that link to the peer document, each link having an anchor
text, and
(iii) identifying a first pattern common to the title of the peer document and the selected anchor text;
(c) identifying a second pattern from the first patterns associated with the plurality of peer documents by identifying the
second pattern from the first patterns associated with the plurality of peer documents based at least in part on the number
of peer documents associated with the first patterns; and
(d) identifying a subject for the target document based on the second pattern and a title of the target document.
|