Page 418 - 《软件学报》2025年第5期

2318 软件学报 2025 年第 36 卷第 5 期

综上所述, 本文提出的 DCGSB 模型综合性能优于其他 10 种对比算法, 并且让网络的拓扑信息和属性信息的
生成均服从幂函数形式的分布且引入节点度刻画网络无标度特性的随机块模型能较好地识别属性网络的同配结
构和异配结构.

6 结 论

本文针对无向属性网络, 提出了一种度修正的属性网络广义随机块模型 DCGSB. 该模型联合了网络拓扑信息
和属性信息, 在对网络拓扑信息建模的同时对节点属性进行建模, 并且引入节点的度刻画网络的无标度特性, 更好
地拟合了真实网络, 进而提高了社团检测精度. 统一的幂函数形式的分布有利于模型参数的估计, 利用 EM 算法具
有较好的算法收敛性. 实验结果显示: 本文的模型 DCGSB 能够发现网络的多种结构, 相较于比较的现有其他算法
在性能上有不同程度的提升, 这说明网络拓扑信息和属性信息的生成均服从幂函数形式的分布并且引入节点的度
可以更好地拟合真实网络, 对社团检测精度也有提升.
本文主要研究了广义随机块模型在无向属性网络中的社团检测问题, 下一步可以考虑将广义随机块模型
GSB 扩展到其他类型的网络. 同时, 本文在模型参数估计时采用了 EM 算法. 但是 EM 算法随着网络规模的增大,
会增加计算量. 在未来工作中, 可以考虑其他参数估计方法.

References:
[1] Zhou Y, Cheng H, Yu JX. Clustering large attributed graphs: An efficient incremental approach. In: Proc. of the 2010 IEEE Int’l Conf. on
Data Mining. Sydney: IEEE, 2010. 689–698. [doi: 10.1109/ICDM.2010.41]
[2] Clauset A, Moore C, Newman MEJ. Hierarchical structure and the prediction of missing links in networks. Nature, 2008, 453(7191):
98–101. [doi: 10.1038/nature06830]
[3] Newman MEJ. Detecting community structure in networks. The European Physical Journal B, 2004, 38(2): 321–330. [doi: 10.1140/epjb/
e2004-00124-y]
[4] Newman MEJ. Fast algorithm for detecting community structure in networks. Physical Review E, 2004, 69(6): 066133. [doi: 10.1103/
PhysRevE.69.066133]
[5] Girvan M, Newman MEJ. Community structure in social and biological networks. Proc. of the National Academy of Sciences of the
United States of America, 2002, 99(12): 7821–7826. [doi: 10.1073/pnas.122653799]
[6] Wang HR, Ding DZ, Zhang M. Gradient-based overlapping hierarchical community detection. Computer Systems & Applications, 2021,
30(8): 207–212 (in Chinese with English abstract). [doi: 10.15888/j.cnki.csa.008016]
[7] Holland PW, Laskey KB, Leinhardt S. Stochastic blockmodels: First steps. Social Networks, 1983, 5(2): 109–137. [doi: 10.1016/0378-
8733(83)90021-7]
[8] Airoldi EM, Blei DM, Fienberg SE, Xing EP. Mixed membership stochastic blockmodels. In: Proc. of the 21st Int’l Conf. on Neural
Information Processing Systems. Vancouver: Curran Associates Inc., 2008. 33–40.
[9] Karrer B, Newman MEJ. Stochastic blockmodels and community structure in networks. Physical Review E, 2011, 83(1): 016107. [doi: 10.
1103/PhysRevE.83.016107]
[10] Lu XY, Szymanski BK. A regularized stochastic block model for the robust community detection in complex networks. Scientific
Reports, 2019, 9(1): 13247. [doi: 10.1038/s41598-019-49580-5]
[11] Shen HW, Cheng XQ, Guo JF. Exploring the structural regularities in networks. Physical Review E, 2011, 84(5): 056111. [doi: 10.1103/
PhysRevE.84.056111]
[12] Jiang JQ. Stochastic block model and exploratory analysis in signed networks. Physical Review E, 2015, 91(6): 062805. [doi: 10.1103/
PhysRevE.91.062805]
[13] Xing EP, Fu WJ, Song L. A state-space mixed membership blockmodel for dynamic network tomography. The Annals of Applied
Statistics, 2010, 4(2): 535–566. [doi: 10.1214/09-AOAS311]
[14] Aicher C, Jacobs AZ, Clauset A. Learning latent block structure in weighted networks. Journal of Complex Networks, 2015, 3(2):
221–248. [doi: 10.1093/comnet/cnu026]
[15] Li Y, Chen HC, Yang B. Reparameterized stochastic block model adaptive to heterogeneous degree and block distributions. IEEE Access,
2018, 6: 37615–37626. [doi: 10.1109/ACCESS.2018.2853115]
[16] Qiao MY, Yu J, Bian W, Li Q, Tao DC. Adapting stochastic block models to power-law degree distributions. IEEE Trans. on