normalize_A: Normalize an affinity matrix
In miguelbiron/threepRop: R Implementation of the 3prop Label Propagation Algorithm

Description Usage Arguments Details Value References Examples

This function provides different modes of normalizing an affinity matrix.

1	normalize_A(A, M_type)

`A`	a symmetric affinity matrix. Can be dense or sparse from package `Matrix`.
`M_type`	a character string with one of "asym", "sym", or "as_is" (see details).

All methods for normalization start by removing possible diagonal elements. The available methods are:

asym: asymmetric normalization, which yields a singly stochastic matrix, where rows add to 1.
sym: symmetric normalization, yielding a symmetric matrix.
rm_diag: only removes possible diagonal elements.

A sparse normalized affinity matrix.

Mostafavi, S., Goldenberg, A., & Morris, Q. (2012). Labeling nodes using three degrees of propagation. PloS one, 7(12), e51947.

1
2
3

# simulate_simple_SBM does not produce self connections
A = as.matrix(simulate_simple_SBM(25L, 0.2, 0.001, 0.25)$A)
all.equal(normalize_A(A, "rm_diag"), A)