Voting Model for Ranking Web Pages

Maxim Lifantsev

Department of Computer Science,
SUNY at Stony Brook,
Stony Brook, NY 11794, U.S.A.
April 2000

maxim@cs.sunysb.edu

Abstract

We describe a new model, the voting model, for computing various kinds of Web page ranks based on the global linkage structure of the Web, to be used by a search engine to estimate quality of Web pages. The voting model has a number of advantages over the random surfer model used in the Google search engine [4]: it subsumes the surfer model providing trivially converging computation of various rank types and better extensibility due to its conceptual simplicity. It also opens up a way to combine the advantages of a human-compiled Web resource directory and an automatic ranking search engine by providing a uniform rank computation method that can utilize both the directory data and the Web linkage graph.

The paper: dvi, ps, or pdf (6 letter pages; as of April 2000)
Slides of the talk presented at IC'00: dvi or ps (9 letter pages; as of June 2000) 

@misc{Lifantsev00:IC,
  author    = "Maxim Lifantsev",
  title     = "Voting Model for Ranking {Web} Pages",
  booktitle = "Proceedings of the International Conference
               on Internet Computing (Las Vegas, Nevada, U.S.A.)",
  editor    = "Peter Graham and Muthucumaru Maheswaran",
  publisher = "CSREA Press",
  address   = "Las Vegas",
  pages     = "143--148",
  year      = "2000",
  month     = "June",
}

/\ Research Papers

Last updated on Dec. 17, 2000 by Maxim Lifantsev
Comments, Suggestions?