Parallel algorithm for query content based webpages ranking

Abstract

The Intelligent Surfer is one of algorithms designed for ranking of search engine results. It is an interesting combination of the PageRank algorithm and the content of web pages. Its main disadvantage is long computation time compared to the PageRank computation time. A computation of the PageRank itself is a very time-consuming process. A lot of papers with topic of efficiency and speed-up of the PageRank computation were published. This paper brings a proposal of speed-up of the Intelligent Surfer algorithm in three steps denoted as the CZDIS algorithm. Experiments with web graph of 1 million nodes size proved that proposed algorithm is usable solution for search engine results ranking in dependance on the page content. Successful implementation of Czech language model experimentally proved possibility of Intelligent Surfer application to different, non English languages. © 2009 Springer Berlin Heidelberg.