Please use this identifier to cite or link to this item:
Title: MISAGA : an algorithm for mining interesting subgraphs in attributed graphs
Authors: He, TT 
Chan, KCC 
Keywords: Attributed graph
Clustering algorithms
Community detection
Graph clustering
Information theoretic measure
Interesting subgraphs
Issue Date: 2018
Publisher: Institute of Electrical and Electronics Engineers
Source: IEEE transactions on cybernetics, May 2018, v. 48, no. 5, p. 1369-1382 How to cite?
Journal: IEEE transactions on cybernetics 
Abstract: An attributed graph contains vertices that are associated with a set of attribute values. Mining clusters or communities, which are interesting subgraphs in the attributed graph is one of the most important tasks of graph analytics. Many problems can be defined as the mining of interesting subgraphs in attributed graphs. Algorithms that discover subgraphs based on predefined topologies cannot be used to tackle these problems. To discover interesting subgraphs in the attributed graph, we propose an algorithm called mining interesting subgraphs in attributed graph algorithm (MISAGA). MISAGA performs its tasks by first using a probabilistic measure to determine whether the strength of association between a pair of attribute values is strong enough to be interesting. Given the interesting pairs of attribute values, then the degree of association is computed for each pair of vertices using an information theoretic measure. Based on the edge structure and degree of association between each pair of vertices, MISAGA identifies interesting subgraphs by formulating it as a constrained optimization problem and solves it by identifying the optimal affiliation of subgraphs for the vertices in the attributed graph. MISAGA has been tested with several large-sized real graphs and is found to be potentially very useful for various applications.
ISSN: 2168-2267
EISSN: 2168-2275
DOI: 10.1109/TCYB.2017.2693558
Appears in Collections:Journal/Magazine Article

View full-text via PolyU eLinks SFX Query
Show full item record


Citations as of Nov 9, 2018


Citations as of Nov 14, 2018

Page view(s)

Citations as of Nov 12, 2018

Google ScholarTM



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.