DocumentCode :
3105756
Title :
Fast Random Walk with Restart and Its Applications
Author :
Tong, Hanghang ; Faloutsos, Christos ; Pan, Jia-Yu
Author_Institution :
Carnegie Mellon Univ., Pittsburg, PA
fYear :
2006
fDate :
18-22 Dec. 2006
Firstpage :
613
Lastpage :
622
Abstract :
How closely related are two nodes in a graph? How to compute this score quickly, on huge, disk-resident, real graphs? Random walk with restart (RWR) provides a good relevance score between two nodes in a weighted graph, and it has been successfully used in numerous settings, like automatic captioning of images, generalizations to the "connection subgraphs", personalized PageRank, and many more. However, the straightforward implementations of RWR do not scale for large graphs, requiring either quadratic space and cubic pre-computation time, or slow response time on queries. We propose fast solutions to this problem. The heart of our approach is to exploit two important properties shared by many real graphs: (a) linear correlations and (b) block- wise, community-like structure. We exploit the linearity by using low-rank matrix approximation, and the community structure by graph partitioning, followed by the Sherman- Morrison lemma for matrix inversion. Experimental results on the Corel image and the DBLP dabasets demonstrate that our proposed methods achieve significant savings over the straightforward implementations: they can save several orders of magnitude in pre-computation and storage cost, and they achieve up to 150x speed up with 90%+ quality preservation.
Keywords :
graph theory; random processes; Sherman-Morrison lemma; connection subgraphs; graph partitioning; low-rank matrix approximation; matrix inversion; random walk with restart; weighted graph; Costs; Delay; Design optimization; Error analysis; Heart; Image storage; Linearity; Niobium; Sparse matrices; Time factors;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Data Mining, 2006. ICDM '06. Sixth International Conference on
Conference_Location :
Hong Kong
ISSN :
1550-4786
Print_ISBN :
0-7695-2701-7
Type :
conf
DOI :
10.1109/ICDM.2006.70
Filename :
4053087
Link To Document :
بازگشت