Numerical Algorithms for Personalized Search in Self-Organizing Information Networks by Kamvar Sep;

Author:Kamvar, Sep; [Kamvar, Sep]
Language: eng
Format: epub
ISBN: 557150
Publisher: Princeton University Press
Published: 2010-08-15T00:00:00+00:00

Find an approximation to the global PageRank vector by weighting the local PageRanks of pages in block J by the personalized BlockRank of J.

Induce the personalization vector over pages from the personalization vector over hosts .

Use this approximation as a start vector for a standard PageRank iteration.

Algorithm 15: Personalized BlockRank Algorithm with Induced Jumps

7.6.3 Experiments

We test this algorithm by computing the personalized PageRank of a random surfer who is a graduate student in linguistics at Stanford. When he bores, he has an 80% probability of jumping to the linguistics host www-linguistics.stanford.edu, and a 20% probability of jumping to the main Stanford host www.stanford.edu. Figure 7.6 shows that the speedup of computing the personalized PageRank for this surfer according to Algorithm 15 shows comparable speedup benefits to standard BlockRank. However, the main benefit is that the local PageRank vectors do not need to be computed at all for Personalized BlockRank. The matrix H is formed from the already computed generic PageRank vector. Therefore, the overhead to computing personalized PageRank vectors using the Personalized BlockRank algorithm is minimal.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.