Enhancing the Performance of Monitoring the DCSBM Using Multivariate Control Charts with Estimated Parameters

Document Type : Research Paper


1 Department of Industrial Engineering, Science and Research Branch, Islamic Azad University, Tehran, Iran.

2 Department of Industrial Engineering, Faculty of Engineering, Shahed University, Tehran, Iran.


Many methods are applied to network surveillance for anomaly detection. Some quality control methods have been developed to monitor several quality characteristics simultaneously in different networks. In our study, we use three multivariate process monitoring techniques such as Hotelling’s T2, MEWMA, and MCUSUM to compare to the prior univariate control charts in the Degree-Corrected Stochastic Block Model (DCSBM), a random network model supporting the degree of each node based on Poisson distribution. By estimating parameters in Phase I from many charts, we apply ARL and SDRL metrics for the performance evaluation of multivariate control charts. The advantage of our method is detecting signals faster than previews ones by simulation and this is useful for defining the suitable method in different types of change. Furthermore, the quality of performance in different multivariate methods is displayed in detecting the shifts in the DCSBM. Finally, MCUSUM shows better performance for monitoring local and global changes than other methods.


  • [1] Sengupta S, Woodall WH. Discussion of “Statistical methods for network surveillance”. Appl Stochastic Models Bus Ind. 2018;34 (4):446-448. doi:1002/asmb.2354.
  • [2] Holme P, Saramäki J. Temporal networks. Phys Rep. 2012;519(3):97-125. doi:1016/j.physrep.2012.03.001.
  • [3] Wilson JD, Stevens NT, Woodall WH. Modeling and detecting change in temporal networks via the degree corrected stochastic block model. Qual Reliab Eng Int. 2019;35(5):1363-1378. doi:1002/qre.2520.
  • [4] Savage D, Zhang X, Yu X, et al. Anomaly detection in online social networks. Soc Networks. 2014;39:62-70. doi:1016/j.socnet.2014.05.002.
  • [5] Woodall WH, Spitzner DJ, Montgomery DC, et al. Using control charts to monitor process and product quality profiles. J Qual Technol. 2004;36(3):309-320. doi:1080/00224065.2004.11980276.
  • [6] Yazdi AA, Hamadani AZ, Amiri A. Phase II monitoring of multivariate simple linear profiles with estimated parameters. J Ind Eng Int. 2019;15(4):557-570. doi:1007/s40092-019-0305-y.
  • [7] Saleh NA, Mahmoud MA, Keefe MJ, et al. The difficulty in designing Shewhart X and X control charts with estimated parameters. J Qual Technol. 2015;47(2):127-138. doi:1080/00224065.2015.11918120.
  • [8] Mahmoud MA, Maravelakis PE. The performance of multivariate CUSUM control charts with estimated parameters. J Stat Comput Simul. 2013;83(4):721-738. doi: 1080/00949655.2011.633910.
  • [9] Javaheri A, Houshmand AA. Average run length comparison of multivariate control charts. J Stat Comput Simul. 2001;69(2):125-140. doi: 1080/00949650108812086.
  • [10] Maleki MR, Castagliola P, Amiri A, et al. The effect of parameter estimation on phase II monitoring of poisson regression profiles. Commun Stat- Simul Comput. 2019;48(7):1964-1978. doi:1080/03610918.2018.1429619.
  • [11] Najafi H, Saghaei A. Statistical monitoring for change detection of interactions between nodes in networks: With a case study in financial interactions network.  Commun Stat- Theory Methods. 2020. doi:1080/03610926.2020.1725830.
  • [12] Woodall WH, Zhao MJ, Paynabar K, et al. An overview and perspective on social network monitoring. IISE Trans. 2017;49(3):354-365. doi:1080/0740817X.2016.1213468.
  • [13] McCulloh I. Detecting changes in a dynamic social network. Institute for Software Research School of Computer Science Carnegie Mellon University; 2009.
  • [14] McCulloh I, Carley KM. Detecting change in longitudinal social networks. Military Academy West Point NY Network Science Center (NSC); 2011. doi:21307/joss-2019-031.
  • [15] Perry MB. An EWMA control chart for categorical processes with applications to social network monitoring. J Qual Technol. 2019. doi:1080/00224065.2019.1571343.
  • [16] Taboran R, Sukparungsee S, Areepong Y. Mixed Moving Average-Exponentially Weighted Moving Average Control Charts for Monitoring of Parameter Change. In: Proceeding of the International MultiConference of Engineers and Computer Scientists. 2019;13-15.
  • [17] Aly AA, Saleh NA, Mahmoud MA. An adaptive EWMA control chart for monitoring zero-inflated Poisson processes. Commun Stat Simul Comput. 2019. doi:1080/03610918.2019.1676437.
  • [18] Aparisi F, Haro CL. A comparison of T 2 control charts with variable sampling schemes as opposed to MEWMA chart. Int J Prod Res. 2003;41(10):2169-2182. doi:1080/0020754031000138655.
  • [19] Aparisi F, Avendaño G, Sanz J. Techniques to interpret T2 control chart signals. IIE Trans. 2006;38(8):647-657. doi:1080/07408170600692200.
  • [20] Mukherjee CS, Peng P, Zhang J. Recovering Unbalanced Communities in the Stochastic Block Model With Application to Clustering with a Faulty Oracle. 2022. arXiv preprint arXiv:2202.08522.
  • [21] Alaluusua K, Leskelä L. Consistent Bayesian community recovery in multilayer networks. 2022. arXiv preprint arXiv:2202.05823.
  • [22] Seif M, Nguyen D, Vullikanti A, Tandon R. Differentially Private Community Detection for Stochastic Block Models. 2022. arXiv preprint arXiv:2202.00636.
  • [23] Qing H, Wang J. Impact of regularization on spectral clustering under the mixed membership stochastic block model. 2021. arXiv preprint arXiv:2107.14705.
  • [24] Vaca-Ramírez F, Peixoto TP. Systematic assessment of the quality of fit of the stochastic block model for empirical networks. 2022. arXiv preprint arXiv:2201.01658.
  • [25] Yu L, Woodall WH, Tsui KL. Detecting node propensity changes in the dynamic degree corrected stochastic block model. Soc Networks. 2018;54:209-227. doi:1016/j.socnet.2018.03.004.
  • [26] Yu L, Zwetsloot IM, Stevens NT, Wilson JD, Tsui KL. Monitoring dynamic networks: A simulation‐based strategy for comparing monitoring methods and a comparative study. Quality and Reliability Engineering International. 2021.
  • [27] Herlau T, Schmidt MN, Mørup M. Infinite-degree-corrected stochastic block model. Phys Rev E. 2014;90 (3):032819. doi:1103/PhysRevE.90.032819.
  • [28] Zhang X, Wang X, Zhao C, et al. Degree-corrected stochastic block models and reliability in networks. Physica A. 2014;393:553-559. doi:1016/j.physa.2013.08.061.
  • [29] Gulikers L, Lelarge M, Massoulié L. An impossibility result for reconstruction in the degree-corrected stochastic block model. Ann Appl Probab. 2018;28(5):3002-3027. doi: 1214/18-AAP1381.
  • [30] Yan X, Shalizi C, Jensen JE, et al. Model selection for degree-corrected block models. J Stat Mech.: Theory Exp. 2014;(5):P05007. doi:1088/1742-5468/2014/05/P05007.
  • [31] Chen Y, Li X, Xu J. Convexified modularity maximization for degree-corrected stochastic block models. Ann Stat. 2018;46(4):1573-1602. doi:1214/17-AOS1595.
  • [32] Zhao Y, Levina E, Zhu J. Consistency of community detection in networks under degree-corrected stochastic block models. Ann Stat. 2012;40(4):2266-2292. doi:1214/12-AOS1036.
  • [33] Gulikers L, Lelarge M, Massoulié L. A spectral method for community detection in moderately sparse degree-corrected stochastic block models. Adv Appl Probab. 2017;49(3):686-721. doi:1017/apr.2017.18.
  • [34] Wang T, Dai W, Jiao P, et al. Identifying influential nodes in dynamic social networks based on degree-corrected stochastic block model. Int J Mod Phys B. 2016;30(16):1650092. doi:1142/S0217979216500922.
  • [35] Fasino D, Tudisco F. The expected adjacency and modularity matrices in the degree corrected stochastic block model. Special Matrices. 2018;6(1):110-121. doi:1515/spma-2018-0010.
  • [36] Lei J, Zhu L. Generic Sample Splitting for Refined Community Recovery in Degree Corrected Stochastic Block Models. Stat Sin. 2017. doi:5705/ss.202015.0279.
  • [37] Ragain S. Community Detection via Discriminant functions for Random Walks in the degree-corrected Stochastic Block Model. 2017.
  • [38] Qian X, Chen Y, Minca A. Clustering Degree-Corrected Stochastic Block Model with Outliers. 2019. arXiv Preprint arXiv:03305.
  • [39] Ma S, Su L, Zhang Y. Determining the Number of Communities in Degree-corrected Stochastic Block Models. J Mach Learn Res. 2021;22(69):1-63. doi:01028.
  • [40] Passino FS, Heard NA, Rubin-Delanchy P. Spectral clustering on spherical coordinates under the degree-corrected stochastic blockmodel. Technometrics. 2022; 1-12. doi: 1080/00401706.2021.2008503.
  • [41] Hosseini SS, Noorossana R. Performance evaluation of EWMA and CUSUM control charts to detect anomalies in social networks using average and standard deviation of degree measures. Qual Reliab Eng Int. 2018;34(4):477-500. doi:1002/qre.2267.
  • [42] Noorossana R, Hosseini SS, Heydarzade A. An overview of dynamic anomaly detection in social networks via control charts. Qual Reliab Eng Int. 2018;34(4):641-648. doi:1002/qre.2278.
  • [43] Mazrae Farahani E, Baradaran Kazemzadeh R. Phase I monitoring of social network with baseline periods using poisson regression. Commun Stat- Theory Methods. 2019;48(2):311-331. doi:1080/03610926.2017.1408836.
  • [44] Fotuhi H, Amiri A, Taheriyoun AR. A novel approach based on multiple correspondence analysis for monitoring social networks with categorical attributed data. J Stat Comput Simul. 2019;89(16):3137-3164. doi:1080/00949655.2019.1657429.
  • [45] Farahani EM, Baradaran Kazemzadeh R, Noorossana R, et al. A statistical approach to social network monitoring. Commun Stat- Theory Methods. 2017;46(22):11272-11288. doi:1080/03610926.2016.1263741.
  • [46] Azarnoush B, Paynabar K, Bekki J, et al. Monitoring temporal homogeneity in attributed network streams. J Qual Technol. 2016;48(1):28-43. doi:1080/00224065.2016.11918149.
  • [47] Miller BA, Arcolano N, Bliss NT. Efficient anomaly detection in dynamic, attributed graphs: Emerging phenomena and big data. In: 2013 IEEE International Conference on Intelligence and Security Informatics. 2013;179-184. doi:1109/ISI.2013.6578815.
  • [48] Fotuhi H, Amiri A, Maleki MR. Phase I monitoring of social networks based on Poisson regression profiles. Qual Reliab Eng Int. 2018;34(4):572-588. doi:1002/qre.2273.
  • [49] Hazrati‐Marangaloo H, Noorossana R. Detecting outbreaks in temporally dependent networks. Qual Reliab Eng Int. 2019;35(6):1753-1765. doi:1002/qre.2473.
  • [50] Zou N, Li J. Modeling and change detection of dynamic network data by a network state space model. IISE Trans. 2017;49(1):45-57. doi:1080/0740817X.2016.1198065.
  • [51] Aiello W, Chung F, Lu L. A random graph model for massive graphs. In: Proceedings of the thirty-second annual ACM symposium on Theory of computing. 2000;171-180. doi:1145/335305.335326.
  • [52] Holland PW, Laskey KB, Leinhardt S. Stochastic blockmodels: First steps. Soc Networks. 1983;5(2):109-137. doi:1016/0378-8733(83)90021-7.
  • [53] Karrer B, Newman ME. Stochastic blockmodels and community structure in networks. Phys Rev E. 2011;83(1):016107. doi:1103/PhysRevE.83.016107.
  • [54] Vanhatalo E, Kulahci M. The effect of autocorrelation on the Hotelling T2 control chart. Qual Reliab Eng Int. 2015;31(8):1779-1796. doi:1002/qre.1717.
  • [55] Mason RL, Tracy ND, Young JC. Decomposition of T 2 for multivariate control chart interpretation. J Qual Technol. 1995;27(2):99-108. doi:1080/00224065.1995.11979573.
  • [56] Santos-Fernández E. Multivariate statistical quality control using R.Vol. 14. Springer Science & Business Media; 2012. doi:1007/978-1-4614-5453-3.