WPR: a Weighted Approach to PageRank

Abstract

The PageRank algorithm which is used by google as a successful ranking algorithm for ranking its results can be interpreted as leveraging the recommendations of all the other page creators on the web, about how important a page is. But it does not take advantage of the recommendations of page visitors. In PageRank, every page creator propagates the importance score of her/his page to its outgoing links uniformly. In this paper a revised version of PageRank called Weighted PageRank (WPR) is proposed in which this uniform propagation is transformed into a weighted one. These weights are assigned to outgoing links based on the average opinion of page visitors about the importance of pages. This opinion is recorded from search engine logs indicating which search results were clicked most. It is demonstrated that our approach is simply applicable without any significant extra time and storage costs compared to PageRank.
Copyright © 2018 Praise Worthy Prize - All rights reserved.