点击次数：

论文类型：期刊论文

发表时间：2008-12-01

发表刊物：Journal of Computational Information Systems

收录刊物：EI、Scopus

卷号：4

期号：6

页面范围：2499-2507

ISSN号：15539105

摘要：Web page ranking plays an important role in modern web information retrieval systems. The key to page ranking is page importance propagation. Google's PageRank adopts a simple method for page importance propagation, in which the importance of each page is propagated only to its direct neighbors through hyperlinks. In this paper, we propose a general framework for importance propagation. Under this framework, every page that is on a directed path to a targeting page has an impact on the importance of the targeting page, and the impact decays with distance following a negative exponent model. PageRank is a special instance of this framework with a propagation distance of 1. By tuning the propagation distance parameter in this framework, the importance of web pages can be propagated more accurately and faster. Experimental results show that, when the propagation distance is 3, the query precision is increased up to 70% and the iterative times are decreased by 78%, compared with PageRank. © 2008 Binary Information Press.