Web Document Classification based on Hyperlinks and Document Semantics.

An algorithm to classify web documents into subsets based on hyperlinks in documents and their content and Representative documents will be identi ed in each subset based on the proposed similarity de nition.

Abstract

. Besides the basic content, a web document also contains a set of hyperlinks pointing to other related documents. Hyperlinks in a document provide much information about its relation with other web documents. By analyzing hyperlinks in documents, inter-relationship among documents can be identified. In this paper, we will propose an algorithm to classify web documents into subsets based on hyperlinks in documents and their content. Representativedocuments will also be identified in each subset based on the proposed similarity definition With the representative document, searching for related documents can be achieved. 1 Introduction The WWW is growing through a decentralized process, and documents in the web are lack of logical organization. Besides, the enormous number of web documents make the manipulation and further operation on web documents difficult. Although the size of web document set is large, we need not analyze all the web documents as a whole. Web documents can...