The hw-rank: an h-index variant for ranking web pages

Abstract

We introduce a novel ranking of search results based on a variant of the h-index for directed information networks such as the Web. The h-index was originally introduced to measure an individual researcher’s scientific output and influence, but here a variant of it is applied to assess the “importance” of web pages. Like PageRank, the “importance” of a page is defined by the “importance” of the pages linking to it. However, unlike the computation of PageRank which involves the whole web graph, computing the h-index for web pages (the hw-rank) is based on a local computation and only the neighbors of the neighbors of the given node are considered. Preliminary results show a strong correlation between ranking with the hw-rank and PageRank, and moreover its computation is simpler and less complex than computation of the PageRank. Further, larger scale experiments are needed in order to assess the applicability of the method.