Efficient discovery of influential nodes for SIS models in social networks

被引:63
作者
Saito, Kazumi [2 ]
Kimura, Masahiro [1 ]
Ohara, Kouzou [3 ]
Motoda, Hiroshi [4 ]
机构
[1] Ryukoku Univ, Dept Elect & Informat, Otsu, Shiga 5202194, Japan
[2] Univ Shizuoka, Sch Adm & Informat, Shizuoka 4228526, Japan
[3] Aoyama Gakuin Univ, Dept Integrated Informat Technol, Coll Sci & Engn, Sagamihara, Kanagawa 2298558, Japan
[4] Osaka Univ, Inst Sci & Ind Res, Osaka 5670047, Japan
关键词
Information diffusion; SIS model; Influence maximization; Pruning method; Burnout method;
D O I
10.1007/s10115-011-0396-2
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
We address the problem of discovering the influential nodes in a social network under the susceptible/infected/susceptible model that allows multiple activation of the same node, by defining two influence maximization problems: final-time and integral-time. We solve this problem by constructing a layered graph from the original network with each layer added on top as the time proceeds and applying the bond percolation with two effective control strategies: pruning and burnout. We experimentally demonstrate that the proposed method gives much better solutions than the conventional methods that are based solely on the notion of centrality using two real-world networks. The pruning is most effective when searching for a single influential node, but burnout is more powerful in searching for multiple nodes which together are influential. We further show that the computational complexity is much smaller than the naive probabilistic simulation both by theory and experiment. The influential nodes discovered are substantially different from those identified by the centrality measures. We further note that the solutions of the two optimization problems are also substantially different, indicating the importance of distinguishing these two problem characteristics and using the right objective function that best suits the task in hand.
引用
收藏
页码:613 / 635
页数:23
相关论文
共 32 条
[1]   Tracking information epidemics in blogspace [J].
Adar, E ;
Adamic, LA .
2005 IEEE/WIC/ACM International Conference on Web Intelligence, Proceedings, 2005, :207-214
[2]  
Agarwal N., 2008, SIGKDD EXPLORATIONS, V10, P18
[3]  
[Anonymous], P 13 INT WORLD WID W
[4]  
[Anonymous], 2007, P 22 NAT C ART INT, DOI DOI 10.5555/1619797.1619865
[5]  
[Anonymous], 2007, P 16 INT C WORLD WID
[6]   Efficient Influence Maximization in Social Networks [J].
Chen, Wei ;
Wang, Yajun ;
Yang, Siyu .
KDD-09: 15TH ACM SIGKDD CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, 2009, :199-207
[7]  
Domingos P, 2005, IEEE INTELL SYST, V20, P80
[8]  
Domingos P., 2001, KDD-2001. Proceedings of the Seventh ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, P57, DOI 10.1145/502512.502525
[9]   Talk of the network: A complex systems look at the underlying process of word-of-mouth [J].
Goldenberg, J ;
Libai, B ;
Muller, E .
MARKETING LETTERS, 2001, 12 (03) :211-223
[10]   ON THE CRITICAL-BEHAVIOR OF THE GENERAL EPIDEMIC PROCESS AND DYNAMICAL PERCOLATION [J].
GRASSBERGER, P .
MATHEMATICAL BIOSCIENCES, 1983, 63 (02) :157-172