RANKING THE PAGES OF THE WORLD WIDE WEB

Abstract

Query search engines are fundamental tools in locating documents satisfying to Web surfers´ interests. A Web search engine enumerates no more than few hundreds of documents for any key word search query. The quality of a search engine largely depends on its ranking algorithm, the heuristics applied for selecting the hit list from the pages containing the key word. This extended abstract discusses general issues about ranking algorithms, then focuses on PageRank, the ranking algorithm implemented in Google. We provide an alternative formulation of PageRank and discuss some further extensions derived from our formulation. Finally, the performance of the modified algorithms is evaluated and compared with other well-known ranking schemes in novel graph based experiments computed on a collection of approximately one million documents downloaded from the Irish domain.