A study of link-competition in a hyperlinked environment

Abstract

In the early 1994, the World Wide Web Worm, one of the early search engines, indexed more than 110,000 web pages that could be accessed on the WWW and provided search services for those pages. At the present time, the Google can index more than 3.3 billion pages and allows users to search the whole pages. Keeping pace with the intense expansion of information, a lot of studies have been researched to improve the quality of search results so that some algorithms capable of ranking web pages according to the concept of links such as HITS and PageRank were introduced. Actually, search engines based on the web page ranking algorithm on the ground of the concept of links show much better performance than others. It is clear that the number of hits to a certain site can severely vary by positions in which a hyperlink to the site is situated and the size of the hyper link. Current text-based web robots can trace the position of certain contents within a file, but cannot within a web browser, which displays those contents. This paper introduces a system to analyze the position and size of hyperlinks displayed in web pages, and represents the analysis result of page views using a web data mining system that can analyze logs in order to survey utilization for certain web pages.