In srdyal/diffusiontest: Analysis of Diffusion and Contagion Processes on Networks

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>"
)

Preliminaries

We are interested in running a large scale simulation featuring hundreds of thousands (if not millions) of vertices. Before we proceed, you need to consider the following:

1. Networks in netdiffuseR are sparse-matrices, thus, as long as the network is sparsely connected, it will fit in a regular computer.
2. At the same time, converting a sparse-matrix into a dense matrix at this scale is not a good idea, as you computer may crash. This would be the case, for example, if you want to calculate the geodesic matrix.
3. Following the previous point, using summary(..., skip.moran = FALSE)--the default behavior--is not a good idea. When dealing with large graphs, set skip.moran = TRUE to avoid memory overflow.

Case 1: Single simulation

Suppose we want to simulate a diffusion process with the following parameters:

• Network type: Small world with parameters $n=200k, k = 10, p = .1$

library(netdiffuseR)

We no proceed with the simulation

ans_sw <- rdiffnet(n = 200, t = 10)

srdyal/diffusiontest documentation built on Sept. 2, 2023, 2:49 p.m.