Speaker
Description
Arguably the most fundamental problem in Network Science is finding structure within a complex network. One approach is to partition the nodes into communities that are more densely connected than one expects in a random network. “The” community structure corresponds to the partition that maximizes a measure that quantifies this idea. Finding the maximizing partition, however, is a computationally difficult NP-complete problem. We explore the use of a recently introduced algorithmic scheme [Guo, Singh, and Bassler, Sci. Rep. 9, 14234 (2019)] to find the structure of a set of benchmark networks. The scheme, known as Reduced Network Extremal Ensemble Learning (RenEEL), creates an ensemble of $k$ partitions and updates the ensemble by replacing its worst member with the best of $k’$ partitions found by analyzing a simplified network. The updating continues until consensus is achieved within the ensemble. Varying the values of $k$ and $k’$, we find that the results obey different classes of extreme value statistics and that increasing $k$ is generally much more effective than increasing $k’$ for finding the best partition. Building upon this exploration, we propose to extend the methodology addressed in [Guo, Singh, and Bassler, J. Phys. Complex. 4 (2023) 025001] to bipartite networks. Introducing a novel metric, bipartite generalized modularity density $Q_{bg}$. This function has a tunable parameter that sets the scale for the typical community found. By varying this parameter hierarchical structure can be found in bipartite networks.
| Academic year | 4th year |
|---|---|
| Research Advisor | Dr. Gemunu Gunaratne, Dr. Kevin E Bassler |
