Linking through time: Memory-enhanced community discovery in temporal networks

Temporal networks present a unique challenge regarding the community discovery task. The inherent dynamism of these systems requires an intricate understanding of memory effects and structural heterogeneity, which are often key drivers of network evolution. This study focuses on Markovian temporal networks and addresses these challenges with an innovative community detection method that introduces a modularity function. We specifically demonstrate how our approach enhances the detectability threshold, thereby improving the effectiveness of community detection in such a dynamic setting. We show that by associating memory directly with nodes' memberships and including it into the modularity expression, we can enhance the detectability threshold compared to scenarios where memory is ignored, thus extending the conditions under which communities can be accurately identified. We validate our approach through extensive numerical simulations, confirming its efficacy in a controlled environment. Additionally, by applying our method to real-world data, we not only demonstrate its practicality and robustness but also reveal its capacity to indirectly tackle additional challenges, such as determining the optimal time window for aggregating data in dynamic graphs. Published by the American Physical Society 2024