DynaMo Analyzation: Dynamic Community Detection by Incrementally Maximizing Modularity Analyzation

Real-world systems are dynamic and complex. Networks can be used to represent these real-world systems. Cluster formation is very important and challenging problems regarding the networks, especially since the real-world networks are dynamic and hard to distinguish communities. DynaMo, which is dynamic community detection by incrementally maximizing modularity analysis, is one of the algorithms that can cluster the dynamic networks. In this paper, the principle and methodology of DynaMo is discussed to introduce the basic background of DynaMo. In order to test the performance of DynaMo, an experiment is conducted to compare the performance of DynaMo with that of the other two algorithms (LWEP & activation) by using several real-world datasets. After the experiment, we can see that different evaluation criteria results in a different performance of these three algorithms. We can conclude that for the NMI, precision, and purity value part, activation outperforms DynaMo and LWEP, while in the recall and Fl value part, DynaMo outperforms activation and LWEP.