Spectral Learning Algorithm Reveals Propagation Capability of Complex Networks

doi:10.1109/TCYB.2018.2861568

CORC > 数学与系统科学研究院 > 中国科学院数学与系统科学研究院 > 系统科学研究所

	Spectral Learning Algorithm Reveals Propagation Capability of Complex Networks
	Xu, Shuang 7; Wang, Pei 1,6; Zhang, Chun-Xia 7; Lu, Jinhu2,3,4,5
刊名	IEEE TRANSACTIONS ON CYBERNETICS
	2019-12-01
卷号	49 期号:12 页码:4253-4261
关键词	Complex network important node influential spreader propagation capability SpectralRank (SR)
ISSN号	2168-2267
DOI	10.1109/TCYB.2018.2861568
英文摘要	In network science and the data mining field, a long-lasting and significant task is to predict the propagation capability of nodes in a complex network. Recently, an increasing number of unsupervised learning algorithms, such as the prominent PageRank (PR) and LeaderRank (LR), have been developed to address this issue. However, in degree uncorrelated networks, this paper finds that PR and LR are actually proportional to in-degree of nodes. As a result, the two algorithms fail to accurately predict the nodes' propagation capability. To overcome the arising drawback, this paper proposes a new iterative algorithm called SpectralRank (SR), in which the nodes' propagation capability is assumed to be proportional to the amount of its neighbors after adding a ground node to the network. Moreover, a weighted SR algorithm is also proposed to further involve a priori information of a node itself. A probabilistic framework is established, which is provided as the theoretical foundation of the proposed algorithms. Simulations of the susceptible-infected-removed model on 32 networks, including directed, undirected, and binary ones, reveal the advantages of the SR-family methods (i.e., weighted and unweighted SR) over PR and LR. When compared with other 11 well-known algorithms, the indices in the SR-family always outperform the others. Therefore, the proposed measures provide new insights on the prediction of the nodes' propagation capability and have great implications in the control of spreading behaviors in complex networks.
资助项目	National Key Research and Development Program of China[2016YFB0800401] ; National Natural Science Foundation of China[61773153] ; National Natural Science Foundation of China[11671317] ; National Natural Science Foundation of China[61621003] ; National Natural Science Foundation of China[61532020] ; National Natural Science Foundation of China[11472290] ; Key Scientific Research Projects in Colleges and Universities of Henan[17A120002] ; Basal Research Fund of Henan University[yqpy20140049]
WOS研究方向	Automation & Control Systems ; Computer Science
语种	英语
出版者	IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
WOS记录号	WOS:000485687200017
内容类型	期刊论文
源URL	[http://ir.amss.ac.cn/handle/2S8OKBNM/35547]
专题	系统科学研究所
通讯作者	Wang, Pei
作者单位	1.Henan Univ, Inst Appl Math, Lab Data Anal Technol, Kaifeng 475004, Peoples R China 2.Chinese Acad Sci, Acad Math & Syst Sci, Beijing 100190, Peoples R China 3.Beihang Univ, Beijing Adv Innovat Ctr Big Data & Brain Machine, Beijing 100083, Peoples R China 4.Beihang Univ, State Key Lab Software Dev Environm, Beijing 100083, Peoples R China 5.Beihang Univ, Sch Automat Sci & Elect Engn, Beijing 100083, Peoples R China 6.Henan Univ, Sch Math & Stat, Kaifeng 475004, Peoples R China 7.Xi An Jiao Tong Univ, Sch Math & Stat, Xian 710049, Shaanxi, Peoples R China
推荐引用方式 GB/T 7714	Xu, Shuang,Wang, Pei,Zhang, Chun-Xia,et al. Spectral Learning Algorithm Reveals Propagation Capability of Complex Networks[J]. IEEE TRANSACTIONS ON CYBERNETICS,2019,49(12):4253-4261.
APA	Xu, Shuang,Wang, Pei,Zhang, Chun-Xia,&Lu, Jinhu.(2019).Spectral Learning Algorithm Reveals Propagation Capability of Complex Networks.IEEE TRANSACTIONS ON CYBERNETICS,49(12),4253-4261.
MLA	Xu, Shuang,et al."Spectral Learning Algorithm Reveals Propagation Capability of Complex Networks".IEEE TRANSACTIONS ON CYBERNETICS 49.12(2019):4253-4261.