Abstract

PageRank is a basic metric used to measure the importance of web pages. It is an iterative method, and computed\nby Power method. Power method converges too slowly for the true structure of web. This issues encourages to find out some\nother iterative methods to accelerate PageRank computation. In this paper, we present a novel method that is combination of\nQuadratic Extrapolation and Gauss-Seidel called QE-GSeidel method which accelerate the convergence of PageRank\nmethod. From numerical experiment we show that QE-GSeidel method takes almost 60-65% less number of iteration and 25-\n30% less CPU time than Power, and takes 50-55% less number of iteration and 15-20% less CPU time than Gauss-Seidel in\nPageRank computation.