Geometries of mixed graphs in complex vector spaces. Hierarchies and clusters in complex networks

Estrada, Ernesto
Mathematical Methods for Natural Phenomena 20, (2025)

We introduce several geometric measures for mixed graphs represented by complex-valued
Hermitian adjacency matrices.We define the communicability functions based on the exponential of the
Hermitian adjacency matrix and define complex-valued position vectors. Then, we define a Euclidean
distance as well as complex, and Euclidean angles between these positions vectors for mixed
graphs. Further we introduce K¨ahler and Hermitian angles between different planes among position
vectors and holomorphic and projection planes, respectively. We find several mathematical relations
and inequalities between all these geometric parameters. To illustrate the usability of some of
these indices in the study of real-world networks we study the K¨ahler angle for finding hierarchies
and detecting hierarchical clusters of vertices in ecological food webs, networks of co-purchasing of
political books, a neuronal network, an Internet trolls network, and a software collaboration graph.
These applications give empirical evidence that the K¨ahler angle contains important information about
the structure of mixed graphs which is relevant for real-world applications in the study of complex
networks.


This web uses cookies for data collection with a statistical purpose. If you continue browsing, it means acceptance of the installation of the same.


More info I agree