GraphChi(C++) Personalized Pagerank

[Sela]

Is there an implementation of PPR in C++ version of GrapgChi?

If not, how would you suggest to approach its implementation using GraphChi.

Thanks

[Aapo Kyrola]

Hi Sela,

unfortunately there is only personalized pagerank in the Java-version, using the DrunkardMob algorithm:

https://github.com/GraphChi/graphchi-java/wiki/Personalized-Pagerank-with-DrunkardMob

If you need to compute PPR for just one user a time, it is easy to modify the basic Pagerank for this purpose.

Aapo

--
You received this message because you are subscribed to the Google Groups "graphchi-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to graphchi-discu...@googlegroups.com.
To post to this group, send email to graphchi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/graphchi-discuss/574ac61e-cd1b-4761-a6e1-af4acf399ed8%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Aapo Kyrola
Ph.D. student, http://www.cs.cmu.edu/~akyrola

GraphChi: Big Data - small machine: http://graphchi.org

twitter: @kyrpov

[Sela]

Thanks, I couldn't find the code where you perform the jump to random vertex (which I could modify to be random out of specific set of vertices)?

[Aapo Kyrola]

Lines 105 - 115 in this:

https://github.com/GraphChi/graphchi-java/blob/master/src/main/java/edu/cmu/graphchi/apps/randomwalks/PersonalizedPageRank.java

Aapo

To view this discussion on the web visit https://groups.google.com/d/msgid/graphchi-discuss/e4c07212-992f-4580-afb2-890d6d65fc73%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.