Convexified modularity maximization for degree-corrected stochastic block models (Q2413602)

In this seminal work the non-asymptotic theoretical guarantees are established for robust approximate clustering and perfect clustering. The main theorem guarantees that employed the convex relaxation approach is strongly consistent in the sense of \textit{Y. Zhao} et al. [Ann. Stat. 40, No. 4, 2266--2292 (2012; Zbl 1257.62095)]. The efficiency of the proposed method and its implementation is illustrated for both synthetic and real-world networks.