rw.c File Reference

Various routines that perform random-walk based operations on graphs stored as gk_csr_t matrices. More...

#include <GKlib.h>

Functions
int	gk_rw_PageRank (gk_csr_t *mat, float lamda, float eps, int max_niter, float *pr)

Detailed Description

Various routines that perform random-walk based operations on graphs stored as gk_csr_t matrices.

Author

George Karypis

Version

$Id: rw.c 11078 2011-11-12 00:20:44Z karypis $ 

Function Documentation

gk_rw_PageRank()

int gk_rw_PageRank	(	gk_csr_t *	mat,
		float	lamda,
		float	eps,
		int	max_niter,
		float *	pr
	)

Computes the (personalized) page-rank of the vertices in a graph.

Parameters

mat	is the matrix storing the graph.
lamda	is the restart probability.
eps	is the error tolerance for convergance.
max_niter	is the maximum number of allowed iterations.
pr	on entry stores the restart distribution of the vertices. This allows for the computation of personalized page-rank scores by appropriately setting that parameter. On return, pr stores the computed page ranks.

Returns

the number of iterations that were performed.