基于时间加权的重叠社区检测算法研究

李慧; 马小平; 张舒; 施珺; 李存华; 仲兆满

doi:10.16383/j.aas.c180559

基于时间加权的重叠社区检测算法研究

doi: 10.16383/j.aas.c180559

李慧^{1, 4,},
马小平^2, ,,
张舒^3,,
施珺^1,,
李存华^1,,
仲兆满^{1, 4,}

1.
江苏海洋大学计算机工程学院连云港 222005
2.
中国矿业大学信息与控制工程学院徐州 221008
3.
江苏海洋大学商学院连云港 222005
4.
江苏省海洋资源开发研究院连云港 222005

基金项目:

国家自然科学基金 61873105

江苏省"333工程" 培养对象, 连云港市第六期"521工程" 培养对象, 江苏省"333工程" 项目 BRA2020261

教育部协同育人项目 201902159041

江苏省教改项目 JGX2019011ZZ

江苏省高等学校自然科学研究项目 19KJB520004

连云港高新区科技计划项目 ZD201912

详细信息

作者简介:
李慧  博士, 江苏海洋大学计算机工程学院副教授. 主要研究方向为个性化推荐, 社会网络分析. E-mail: shufanzs@126.com

张舒  江苏海洋大学商学院讲师. 主要研究方向为智能信息处理. E-mail: lih@cumt.edu.cn

施珺  江苏海洋大学计算机工程学院教授. 主要研究方向为智能信息处理. E-mail: sj_lfg@hotmail.com

李存华  博士, 江苏海洋大学计算机工程学院教授. 主要研究方向为数据挖掘. E-mail: cli2000@126.com

仲兆满  博士, 江苏海洋大学计算机工程学院副教授. 主要研究方向为中文信息处理. E-mail: zhongzhaoman@163.com

通讯作者:
马小平博士, 中国矿业大学信电学院教授. 主要研究方向为智能计算. 本文通信作者. E-mail: xpma@cumt.edu.cn

计量
- 文章访问数: 1227
- HTML全文浏览量: 317
- PDF下载量: 169
- 被引次数: 0
出版历程
- 收稿日期: 2018-10-30
- 录用日期: 2019-05-19
- 刊出日期: 2021-04-01

Research of Overlap Community Detection Algorithm Based on Time-Weighted

LI Hui^{1, 4
,},
MA Xiao-Ping^{2
, ,},
ZHANG Shu^3
,,
SHI Jun^1
,,
LI Cun-Hua^1
,,
ZHONG Zhao-Man^{1, 4
,}

1.
School of Computer Engineering, Jiangsu Ocean University, Lianyungang 222005
2.
School of Information and Control Engineering, China University of Mining and Technology, Xuzhou 221008
3.
School of Business, Jiangsu Ocean University, Lianyungang 222005
4.
Jiangsu Institute of Marine Resources Development, Lianyungang 222005

Funds:

National Natural Science Foundation of China 61873105

"333 Project" cultivation object of Jiangsu province, "521 Project" Cultivation Object of Lianyungang, "333 Project" of Jiangsu Province BRA2020261

Ministry of Education Collaborative Education Project 201902159041

Education Reform Project of Jiangsu Province JGX2019011ZZ

Natural Science Research Project of Higher Education of Jiangsu Province 19KJB520004

Lianyungang High-tech Zone Science and Technology Project ZD201912

More Information

Author Bio:
LI Hui  Ph. D., associate professor at the School of Computer Engineering, Jiangsu Ocean University. Her research interest covers personality recommendation and social network analysis

ZHANG Shu  Lecturer at the School of Business, Jiangsu Ocean University. Her main research interest is information processing

SHI Jun  Professor in the Department of Computer Science, Jiangsu Ocean University. Her main research interest is information processing

LI Cun-Hua  Ph. D., professor in the Department of Computer Science, Jiangsu Ocean University. His main research interest is data mining

ZHONG Zhao-Man  Ph. D., associate professor in the Department of Computer Science, Jiangsu Ocean University. His main research interest is Chinese information processing

Corresponding author: MA Xiao-Ping Ph. D., professor at the School of Information and Electrical Engineering, China University of Mining & Technology. His main research interest is intelligent computing. Corresponding author of this paper

摘要

摘要: 随着网络结构的不断扩大和日益复杂, 重叠社区发现技术对挖掘复杂网络深层潜在结构具有重要意义. 本文提出一种基于时间加权的重叠社区检测算法. 该方法考虑了用户兴趣的时间因素, 构建带有时间加权链接的用户-用户图. 接着, 基于网络节点的影响力计算用户全局相似度, 在此基础上通过计算节点的中心度作为度量节点对社区结构影响力的重要性指标, 从而提出一种社区中心点的选取方法. 最后, 通过效用函数的迭代计算实现重叠社区检测. 利用人工网络和真实网络对提出的算法进行验证, 实验结果表明: 相对于传统的社区发现方法, 该算法在社区发现质量和计算效率方面都优于许多已有重叠社区发现算法.
- 社会网络 /
- 加权 /
- 重叠 /
- 社区 /
- 检测
Abstract: With the continuous expansion and complexity of network structure, overlapping community discovery technology is of great signiflcance to excavate the deep potential structure of complex network. This article presents an overlapping community detection algorithm based on time time-weighted. Considering the time factor of user interest, this method constructs a user-user graph with time-weighted links. Then, the global similarity of users is calculated based on the influence of network nodes. On this basis, the centrality of nodes is calculated as an important index to measure the impact of nodes on community structure, and a method to select community centers is proposed. Finally, overlapping community detection is realized by iteration of utility function. The proposed algorithm is validated by artiflcial network and real network. The experimental results show that compared with traditional community discovery methods, the proposed algorithm outperforms many existing overlapping community discovery algorithms in terms of community discovery quality and computational efficiency.
- Social network /
- time-weighted /
- overlap /
- community /
- detection
Recommended by Associate Editor DUAN Shu-Kai
注释:

1) 本文责任编委段书凯

HTML全文

图 1 Polblogs数据集中参数α对Q值的影响结果

Fig. 1 The influence of different α on the Q in Polblogs data set

下载: 全尺寸图片幻灯片

图 2 DBLP数据集中参数α对Q值的影响结果

Fig. 2 The influence of different α on the Q in DBLP data set

下载: 全尺寸图片幻灯片

图 3 重叠社区的模块度随社区数量的变化情况

Fig. 3 The influence of different community number on the K

下载: 全尺寸图片幻灯片

图 4 社区数量随阈值β的变化情况

Fig. 4 The influence of different β on different community number K

下载: 全尺寸图片幻灯片

图 5 真实数据集上各算法性能对比实验

Fig. 5 Comparison results of different algorithms on real networks

下载: 全尺寸图片幻灯片

图 6 不同算法运行时间比较

Fig. 6 Execution time comparison of different algorithms

下载: 全尺寸图片幻灯片

表 1 LFR基准网络生成参数说明

Table 1 Parameter setting of LFR benchmark network generation

参数	说明
N	网络的节点数目
k	网络中节点的平均度数
C_min	最小社区的节点数目
C_max	最大社区的节点数目
on	重叠节点的个数
om	重叠节点所从属的社区个数
mu	社区混合参数

下载: 导出CSV

表 2 人工网络数据集

Table 2 Artificial network datasets

编号	N	k	C_min	C_max	mu	on	om
S₁	10 000	20	50	100	0.1~0.7	1 000	3
S₂	100 000	20	100	200	0.1~0.7	5 000	2
S₃	10 000	20	100	200	0.1	1 000	2~7

下载: 导出CSV

表 3 真实数据集

Table 3 Real datasets

编号	名称	节点数	边数	平均度
R1	Karate	34	78	4.75
R2	Doplphins	62	159	5.13
R3	Polbooks	105	441	8.40
R4	Football	115	613	10.66
R5	Folbogs	1 490	16 715	22.44
R6	DBLP	4 000	8 301	2.52

下载: 导出CSV

表 4 mu在不同取值时各算法在人工网络S1上的NMI实验结果

Table 4 NMI experimental results of different algorithms on S1 under different mu value

mu	0.1	0.2	0.3	0.4	0.5	0.6	0.7
TWOCD	0.53	0.45	0.34	0.24	0.23	0.22	0.22
COPRA	0.85	0.79	0.71	0.62	0.42	0.22	0.22
CPM	0.82	0.8	0.62	0.48	0.21	0.22	0.22
LFM	0.52	0.42	0.35	0.22	0.22	0.22	0.22

下载: 导出CSV

表 5 om在不同取值时各算法在人工网络S2上的NMI实验结果

Table 5 NMI experimental results of different algorithms on S2 under different om value

om	2	3	4	5	6	7	8
TWOCD	0.92	0.95	0.88	0.84	0.78	0.72	0.75
COPRA	0.93	0.9	0.83	0.78	0.71	0.65	0.6
CPM	0.92	0.87	0.81	0.78	0.69	0.62	0.62
LFM	0.76	0.68	0.66	0.67	0.65	0.66	0.5

下载: 导出CSV

表 6 on在不同取值时各算法在人工网络S3上的NMI实验结果

Table 6 NMI experimental results of different algorithms on S3 under different on value

on	1 000	2 000	3 000	4 000	5 000	6 000	7 000
TWOCD	0.95	0.95	0.89	0.76	0.67	0.56	0.38
COPRA	0.89	0.83	0.78	0.58	0.32	0.21	0.21
CPM	0.82	0.84	0.79	0.7	0.6	0.47	0.28
LFM	0.43	0.33	0.23	0.23	0.24	0.24	0.24

下载: 导出CSV

表 7 真实数据集上各算法在不同参数取值下性能对比结果

Table 7 Comparison results of different algorithms on different parameter in real networks

Data set	LFM		COPRA		CPM		TWOCD
Data set	α	Q	v	Q	k	Q	α, β	Q
Karate	0.8	0.813	2	0.825	2	0.823	0.4, 0.8	0.8463
Football	1.1	0.645	2	0.656	4	0.707	0.4, 0.8	0.7246
Doplphins	0.8	0.812	6	0.821	5	0.924	0.3, 0.7	0.9005
Polbooks	0.9	0.634	8	0.717	8	0.795	0.4, 0.9	0.7342
Folbogs	1.4	0.122	2	0.466	6	0.625	0.4, 0.8	0.6257
DBLP	0.8	0.787	9	0.745	3	0.797	0.4, 0.8	0.8143

下载: 导出CSV

参考文献(28)

[1]	Newman M E J. The Structure and function of complex networks. SIAM Review, 2003, 45(2): 167-256 doi: 10.1137/S003614450342480
[2]	潘剑飞, 董一鸿, 陈华辉, 等. 基于结构紧密性的重叠社区发现算法. 电子学报, 2019, 47(001): 145-152 https://www.cnki.com.cn/Article/CJFDTOTAL-DZXU201901019.htm Pan Jian-Fei, Dong Yi-Hong, Chen Hua-Hui, et al. The overlapping community discovery algorithm based on compact structure. Acta Electronic Sinica, 2019, 47(001): 145-152 https://www.cnki.com.cn/Article/CJFDTOTAL-DZXU201901019.htm
[3]	李慧, 马小平, 施珺, 李存华, 仲兆满, 蔡虹. 复杂网络环境下基于信任传递的推荐模型研究. 自动化学报, 2018, 44(2): 363-376 doi: 10.16383/j.aas.2018.c160395 Li Hui, Ma Xiao-Ping, Shi Jun, Li Cun-Hua, Zhong Zhao-man, Cai Hong. A recommendation model by means of trust transition in complex network environment. Acta Automatica Sinica, 2018, 44(2): 363-376 doi: 10.16383/j.aas.2018.c160395
[4]	Xu M, Li Y, Li R, et al. EADP: An extended adaptive density peaks clustering for overlapping community detection in social networks. Neurocomputing, 2019, 337(14): 287-302 http://www.sciencedirect.com/science/article/pii/S092523121930102X
[5]	Gergely P, Imre D, Illés F, Tamás V. Uncovering the overlapping community structure of complex networks in nature and society. Nature, 2005, 435: 814-818 doi: 10.1038/nature03607
[6]	Farkas I, ábel D, Palla G, et al. Weighted network modules. New Journal of Physics, 2007, 9(6): 180 doi: 10.1088/1367-2630/9/6/180
[7]	Zhang Z, Wang Z. Mining overlapping and hierarchical communities in complex networks. Physica A: Statistical Mechanics and Its Applications, 2015, 421: 25-33 doi: 10.1016/j.physa.2014.11.023
[8]	Wang L. Using the relationship of shared neighbors to find hierarchical overlapping communities for effective connectivity in IoT. In: Proceedings of the 6th International Conference on Pervasive Computing and Applications. New York, USA: IEEE, 2011. 400-406
[9]	陈俊宇, 周刚, 南煜, 曾琦. 一种半监督的局部扩展式重叠社区发现方法. 计算机研究与发展, 2016, 53(6): 1376-1388 https://www.cnki.com.cn/Article/CJFDTOTAL-JFYZ201606019.htm Chen Jun-Yu, Zhou Gang, Nan Yu, Zeng Qi. Semi-supervised local expansion method for overlapping community detection. Journal of Computer Research and Development, 2016, 53(6): 1376-1388 https://www.cnki.com.cn/Article/CJFDTOTAL-JFYZ201606019.htm
[10]	Lancichinetti A, Fortunato S, Kertész J. Detecting the overlapping and hierarchical community structure in complex networks. New Journal of Physics, 2009, 11(3): 15-33 http://scitation.aip.org/getabs/servlet/GetabsServlet?prog=normal&id=VIRT01000019000015000092000001&idtype=cvips&gifs=Yes
[11]	Murphy R J L. Reliability of marking in eight GCE examinations. British Journal of Educational Psychology, 2011, 48(2): 196-200 doi: 10.1111/j.2044-8279.1978.tb02385.x/abstract
[12]	Lancichinetti A, Radicchi F, Ramasco J J, et al. Finding statistically significant communities in networks. PloS One, 2011, 6(4): e18961 doi: 10.1371/journal.pone.0018961
[13]	Yang J, Zhang X D. Finding overlapping communities using seed set. Physica A: Statistical Mechanics and Its Applications, 2017, 467(1): 96-106
[14]	Su Y J, LeeC C. Overlapping community detection with seed set expansion by local cluster coefficient. In: Proceedings of the International Conference on Consumer Electronics Piscataway, New York, USA: IEEE, 2017. 73-74
[15]	Gregory S. Fuzzy overlapping communities in networks. Journal of Statistical Mechanics-Theory and Experiment, 2011, 2: P02017
[16]	Eustace J, Wang X, Cui Y. Overlapping community detection using neighborhood ratio matrix. Physica A: Statistical Mechanics and its Applications, 2015, 421: 510-521 doi: 10.1016/j.physa.2014.11.039
[17]	Raj E D, Babu L D D. A fuzzy adaptive resonance theory inspired overlapping community detection method for online social networks. Knowledge-Based Systems, 2016, 112(1): 75-87 doi: 10.5555/3045900.3050946
[18]	Su Jianhan, Havens T C. Quadratic program-based modularity maximization for fuzzy community deterction in social network. IEEE Trans. on Fuzzy Systems, 2015, 23(5): 1356-1371 doi: 10.1109/TFUZZ.2014.2360723
[19]	Javed M A, Younis M S, Latif S, et al. Community detection in networks: A multidisciplinary review. Journal of Network and Computer Applications, 2018, 108(15): 87-111 http://www.sciencedirect.com/science/article/pii/S1084804518300560
[20]	Chen Naiyue, Liu Yun, Chao H C. Overlapping community detection using non-negative matrix factorization with orthogonal and sparseness constraints. IEEE Access, 2018, 6: 21266-21274 doi: 10.1109/ACCESS.2017.2783542
[21]	Ahn Y Y, Lehmann S, Bagrow J, et al. Hierarchical link clustering in complex networks. American Physical Society, 2009, 3(2009): 1-10. http://meetings.aps.org/link/baps.2009.mar.v9.1
[22]	Shi C, Cai Y, Fu D, et al. A link clustering based overlapping community detection algorithm. Data & Knowledge Engineering, 2013, 87: 394-404 http://ieeexplore.ieee.org/document/6574677/
[23]	Lim S, Ryu S, Kwon S, et al. LinkSCAN: Overlapping community detection using the link-space transformation. In: Proceedings of the 30th International Conference on Data Engineering. New York, USA: IEEE, 2014. 292-303
[24]	Li M M, Liu J. A link clustering based memetic algorithm for overlapping community detection. Physica A: Statistical Mechanics and Its Application, 2018, 503(1): 410-423 http://www.sciencedirect.com/science/article/pii/S037843711830253X
[25]	Gregory S. Finding overlapping communities in networks by label propagation. New Journal of Physics, 2009, 12(10): 2011-2024 http://arxiv.org/abs/0910.5516v3
[26]	Xie J, Szymanski B K. Towards linear time overlapping community detection in social networks. In: Proceedings of the Pacific-Asia Conference on Knowledge Discovery and Data Mining, New York, USA: IEEE, 2012. 25-36
[27]	Gaiter C, Chen M M, Szymanski B, et al. Identifying robust communities and multi-community nodes by combining top-down and bottom-up approaches to clustering. Scientific Reports, 2015, 5: 16361 doi: 10.1038/srep16361
[28]	Brin S, Page L. Reprint of: The anatomy of a large-scale hypertextual web search engine. Computer Networks, 2012, 56(18): 3825-3833 doi: 10.1016/j.comnet.2012.10.007