US 8,046,361 B2Granted

System and method for classifying tags of content using a hyperlinked corpus of classified web pages

AssigneeYahoo! Inc.

InventorsBörkur Sigurbjörnsson, Roelof van Zwol, Simon E. Overell

Life of this patent

Estimating the term…

Loading patent…

Abstract

An improved system and method for classifying tags of content using a hyperlinked corpus of classified web pages is provided. An anchor text index may be searched to find anchor texts that may match text of the tag, documents referenced by the matching anchor texts may be found, and the documents referenced by the matching anchor texts may be grouped to disambiguate multiple classifications that result from matching the anchor texts with the categories of the reference documents. To resolve ambiguity between multiple classifications, weighted classifications may be used where each document may be assigned a positive weight for a mapping to a category to indicate the confidence of the classification of the document to the category. The classification for the grouping of the documents referenced by the matching anchor texts with greatest frequency may be selected and output as the classification for the tag.

Claims18

Dependent claims

A computer system for classifying a tag associated with content, comprising:

a tag classification engine for classifying a tag associated with content of a web document with a category associated with one or more documents in a classified corpus of hyperlinked web documents referred by one or more anchor texts matching the text of the tag;
a storage operably coupled to the tag classification engine for storing a plurality of categories of tags classified with the category associated with one or more documents in the classified corpus of hyperlinked web documents referred by the one or more anchor texts matching the text of the tag, said storage comprising the classified corpus of hyperlinked web documents;
an anchor text index searchable to find anchor texts in the classified corpus of hyperlinked web documents that match text of the tag;
and
an anchor text search engine operably coupled to the tag classification engine configured to perform:
searching the anchor text index to find the one or more anchor texts in the classified corpus of hyperlinked web documents that match the text of the tag, in order to disambiguate multiple classifications that result from matching the anchor texts with the categories of the reference documents in the classified corpus of hyperlinked web documents;
reducing ambiguity by:
clustering all documents with a same classification;
and
excluding documents accounting for less than five percent of total mappings from a specific anchor.

7 more elements

A computer system for classifying a tag associated with content, comprising:

a computer processing device performing steps of:
searching an anchor text index to find one or more anchor texts in a classified corpus of hyperlinked web documents;
matching text of a tag associated with content of a web document with the one or more anchor texts in the classified corpus of hyperlinked web documents, in order to disambiguate multiple classifications that result from matching the anchor texts with the categories of the reference documents in the classified corpus of hyperlinked web documents wherein ambiguity is reduced by clustering all documents with the same classification;
and
matching the one or more anchor texts with one or more categories of one or more documents in the classified corpus of hyperlinked web documents;
reducing ambiguity by:
clustering all documents with a same classification;
and
excluding documents accounting for less than five percent of total mappings from a specific anchor;
and
an output device operably coupled with the computer processing device, said output device outputting at least one classification for the tag associated with the content of the web document.

8 more elements

Dependencies are inferred from the claim text.