The Distribution of PageRank Follows a Power-Law only for Particular Values of the Damping Factor

• Luca Becchetti, Universita di Roma "La Sapienza", Italy
• Carlos Castillo, Universita di Roma "La Sapienza", Italy

Full text:

Track: Posters

We show that the empirical distribution of the PageRank values in a large set of Web pages does not follow a power-law except for some particular choices of the damping factor. We argue that for a graph with an in-degree distribution following a power-law with exponent between 2.1 and 2.2, choosing a damping factor around 0.85 for PageRank yields a power-law distribution of its values. We suggest that power-law distributions of PageRank in Web graphs have been observed because the typical damping factor used in practice is between 0.85 and 0.90.

Other items being presented by these speakers

• Testing Google Interfaces Modified for the Blind (Posters Track)
• Relationship Between Web Links and Trade (Posters Track)