TY - JOUR

T1 - SBV-Cut

T2 - Vertex-cut based graph partitioning using structural balance vertices

AU - Kim, Mijung

AU - Candan, Kasim

N1 - Funding Information:
K. Selcuk Candan is a Professor of Computer Science and Engineering at the School of Computing, Informatics, and Decision Science Engineering at the Arizona State University. He joined the department in August 1997, after receiving his Ph.D. from the Computer Science Department at the University of Maryland at College Park. Prof. Candan's primary research interest is in the area of management of non-traditional, heterogeneous, and imprecise (such as multimedia, web, and scientific) data. His various research projects in this domain are funded by diverse sources, including the National Science Foundation, Department of Defense, Mellon Foundation, and DES/RSA (Rehabilitation Services Administration). He has published over 140 articles and many book chapters. He has also authored 9 patents. Recently, he co-authored a book titled “Data Management for Multimedia Retrieval” for the Cambridge University Press. Prof. Candan served as an editorial board member of one of the most respected database journals, the Very Large Databases (VLDB) journal. He is also in the editorial board of the Journal of Multimedia. He has served in the organization and program committees of various conferences. In 2006, he served as an organization committee member for SIGMOD'06, the flagship database conference of the ACM and one of the best conferences in the area of management of data. In 2008, he served as a PC Chair for another leading, flagship conference of the ACM, this time focusing on multimedia research (MM'08). More recently, he served as a program committee group leader for ACM SIGMOD'10. In 2011, he will serve as a general co-chair for the ACM MM'11 conference. In 2012 he will serve as a general co-chair for ACM SIGMOD'12.

PY - 2012/2

Y1 - 2012/2

N2 - Graphs are used for modeling a large spectrum of data from the web, to social connections between individuals, to concept maps and ontologies. As the number and complexities of graph based applications increase, rendering these graphs more compact, easier to understand, and navigate through are becoming crucial tasks. One approach to graph simplification is to partition the graph into smaller parts, so that instead of the whole graph, the partitions and their inter-connections need to be considered. Common approaches to graph partitioning involve identifying sets of edges (or edge-cuts) or vertices (or vertex-cuts) whose removal partitions the graph into the target number of disconnected components. While edge-cuts result in partitions that are vertex disjoint, in vertex-cuts the data vertices can serve as bridges between the resulting data partitions; consequently, vertex-cut based approaches are especially suitable when the vertices on the vertex-cut will be replicated on all relevant partitions. A significant challenge in vertex-cut based partitioning, however, is ensuring the balance of the resulting partitions while simultaneously minimizing the number of vertices that are cut (and thus replicated). In this paper, we propose a SBV-Cut algorithm which identifies a set of balance vertices that can be used to effectively and efficiently bisect a directed graph. The graph can then be further partitioned by a recursive application of structurally-balanced cuts to obtain a hierarchical partitioning of the graph. Experiments show that SBV-Cut provides better vertex-cut based expansion and modularity scores than its competitors and works several orders more efficiently than constraint-minimization based approaches.

AB - Graphs are used for modeling a large spectrum of data from the web, to social connections between individuals, to concept maps and ontologies. As the number and complexities of graph based applications increase, rendering these graphs more compact, easier to understand, and navigate through are becoming crucial tasks. One approach to graph simplification is to partition the graph into smaller parts, so that instead of the whole graph, the partitions and their inter-connections need to be considered. Common approaches to graph partitioning involve identifying sets of edges (or edge-cuts) or vertices (or vertex-cuts) whose removal partitions the graph into the target number of disconnected components. While edge-cuts result in partitions that are vertex disjoint, in vertex-cuts the data vertices can serve as bridges between the resulting data partitions; consequently, vertex-cut based approaches are especially suitable when the vertices on the vertex-cut will be replicated on all relevant partitions. A significant challenge in vertex-cut based partitioning, however, is ensuring the balance of the resulting partitions while simultaneously minimizing the number of vertices that are cut (and thus replicated). In this paper, we propose a SBV-Cut algorithm which identifies a set of balance vertices that can be used to effectively and efficiently bisect a directed graph. The graph can then be further partitioned by a recursive application of structurally-balanced cuts to obtain a hierarchical partitioning of the graph. Experiments show that SBV-Cut provides better vertex-cut based expansion and modularity scores than its competitors and works several orders more efficiently than constraint-minimization based approaches.

KW - Clustering

KW - Graph partitioning

KW - Mining methods and algorithms

KW - Vertex-cut

UR - http://www.scopus.com/inward/record.url?scp=84855238538&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84855238538&partnerID=8YFLogxK

U2 - 10.1016/j.datak.2011.11.004

DO - 10.1016/j.datak.2011.11.004

M3 - Article

AN - SCOPUS:84855238538

SN - 0169-023X

VL - 72

SP - 285

EP - 303

JO - Data and Knowledge Engineering

JF - Data and Knowledge Engineering

ER -