Web Mining in Cloud Computing Framework RANKING

Abstract

The dimension of World Wide Web (The Internet) is in billions in terms of web pages and increasing rapidly. With the diversity of web pages available on the web, the high degree relevant information retrieval becomes a major issue. Such huge number of pages not only make the computation complex but also raises the issues of fault tolerance and time complexity. Computing ranking for such large number of web graph on a particular system, makes it prone to system failure and time taking. The present work proposes a distributed ranking system to attain fault tolerance and speedy calculation of Pagerank vector. The computation of rank vector is performed by implementing Pagerank on Mapreduce framework. The pagerank vector is calculated via spectral analysis to make the computation even faster and the results are compared to traditional pagerank scores.