Complex non-backtracking matrix for directed graphs

Sando, Keishi, Hino, Hideitsu

arXiv.org Machine Learning 

In network analysis, various matrix representations have been developed to investigate the structural properties of the corresponding network, among which the non-backtracking (NBT) matrix is one such representation. The NBT matrix is well-known for its relationship with the Ihara zeta function [15] defined as an infinite product over equivalence classes of primitive cycles. It was shown in [15, 36] that, for regular graphs, the reciprocal of the Ihara zeta function can be expressed as a polynomial related to the adjacency matrix. The relation between the zeta function and the determinant of the NBT matrix was elucidated in [12], extended to irregular graphs in [3], and an elementary proof was provided in [35]. The connection between the polynomial and the determinant of the NBT matrix, via the Ihara zeta function, is known as the Ihara's formula.

Duplicate Docs Excel Report

Title
None found

Similar Docs  Excel Report  more

TitleSimilaritySource
None found