Method of visualizing homepages using certain quantifiers and its verification
O. Yamagata, N. Kataoka, K. Tsuda, M. Morisue · 2002
The paper proposes the concept of "page-interest-index", which is an extraction of the characteristics of homepage information according to the degree to which the user's interest is satisfied. The proposed page-interest-index is a value that correlates such characteristics of homepage information with the degree of satisfaction of the user's interest. The concepts of "homepage-character" and "browsing-character" are also defined. They are the values consisting of a sequence of page-interest-indices in the order of pages browsed by the user. They are useful for getting an overview of the homepage. We have developed a verification system and applied it to homepages of consumer products. It has been observed that the homepage-character calculated from the Web documents takes an "upward convex" shape. Meanwhile, we have analyzed the probability of transition between pages, which were obtained from the Web access log, as a hidden Markov model (HMM). We confirmed that the page-interest-index, homepage-character and browsing-character proposed in the paper are useful quantifiers of homepages for making a visual representation of homepage information.