Extraction of leader-pages in WWW: An Improved Approach based on Artificial Link Based Similarity and Higher-Order Link Based Cyclic Relationships

摘要

WWW is the most popular and interactive medium to disseminate information. It creates many new challenges. Several initiatives have been taken to extract different kinds of knowledge from web. In our paper "Leader-page Resources in the World Wide Web" we defined a new method to rank the web pages entirely using the hyperlink information. The notion of "Leader-Page" is extended from the concept of leader from the leadership theory and the social networks. In a community, a leader is a person who interacts the most with other members of the community and whose characteristics are most similar to the characteristics of other members of the community. We have extended these properties of leader to identify leader-pages in WWW. In this paper we propose an improved approach to measure the "leadership score" of a web page based on artificial link based similarity and higher-order link based cyclic relationships it establishes with other web pages of the cyber community.