Is it possible to normalize a symmetric matrix without breaking symmetry?

1.1k Views Asked by At

To add a bit more detail to the question in the title, I have an adjacency matrix of a graph, which is symmetric. Symmetric adjacent matrices have nice theoretical properties. However, they break down when I naively normalize the matrix (e.g. so that each row sum to 1), since it's no longer symmetric. Are there ways to get around this issue?