An Empirical Investigation of PageRank and Its Variants in Ranking Pages on the Web

Web Information Retrieval (IR) has been successful with page-ranking algorithms that order web pages based on their rankings and relevance. These ranking algorithms are one of the success factors behind today's popular web search engines including Ask, Bing, Google, and Yahoo! etc., with Google on the top since long. Besides other ranking signals, Google uses PageRank algorithm in ranking the search results, which makes Google successful and superior to others. Since its inception in 1998, it has been at the heart of Google's ranking system and considered a serious breakthrough in ranking web pages on the Web. Researchers and scientists followed this link-based strategy and came up with similar ranking algorithms including Weighted PageRank, which together with PageRank, have been the focus of research articles covering several aspects and properties. In this article, we report an empirical investigation of PageRank algorithm and Weighted PageRank algorithm with respect to the property of convergence. Results of the study show that both these algorithms are limited especially with respect to convergence. Based on these results, we propose a new flavor of PageRank called Ratio-based Weighted PageRank that performs better than PageRank and Weighted PageRank algorithms especially in terms of convergence.