Node seniority ranking

Recent advances in graph theory suggest that is possible to identify the oldest nodes of a network using only the graph topology. Here we report on applications to heterogeneous real world networks. To this end, and in order to gain new insights, we propose the theoretical framework of the Estrada communicability. We apply it to two technological networks (an underground, the diffusion of a software worm in a LAN) and to a third network representing a cholera outbreak. In spite of errors introduced in the adjacency matrix of their graphs, the identification of the oldest nodes is feasible, within a small margin of error, and extremely simple. Utilizations include the search of the initial disease-spreader (patient zero problem), rumors in social networks, malware in computer networks, triggering events in blackouts, oldest urban sites recognition.

💡 Research Summary

The paper addresses the long‑standing problem of identifying the oldest nodes—those that were present at the earliest stages of a network’s formation—using only the network’s topology. Traditional centrality measures (degree, betweenness, eigenvector, etc.) capture static importance but ignore temporal order, making them unsuitable for “patient‑zero” type investigations. To overcome this limitation, the authors introduce a theoretical framework based on Estrada communicability (EC).

EC is defined as the diagonal entries of the matrix exponential of the adjacency matrix, exp(A). Mathematically, if A is the N × N adjacency matrix, then exp(A)=∑_{k=0}^{∞} A^{k}/k! and the i‑th EC value is EC_i =