A Clickstream Based Web Page Importance Metric for Customized Search Engines

The immense growing dimension of the World Wide Web causes considerable obstacles for search engine applications. Since the performance of any Web search engine regarding the degree of result set’s accuracy and the portion of the presence of authoritative Web pages in result set is highly dependent to the applied Web page importance metric, therefore any enhancement to the existing metrics or designing novel related algorithms could guarantee better outcomes of search engine applications. Regarding the fact that employing the existing link dependent Web page importance metrics in search engines is not an absolute solution because of their accurateness dependency to the downloaded portion of the Web and their incapability in covering authoritative dark Web pages, therefore proposing and discussing on link independent approaches could be a solution to the mentioned barriers. This paper reviews our clickstream based Web page importance metric of

LogRank

that is independent of the link structure of the Web graph for extracting the best result set from the specific Web domain boundary and for importance estimation of a whole Web domain. Moreover, our Web page classification approach in order to be used in Web site importance calculation will be reviewed.