1

Im stuck with the random walk. I tried to simulate the random on a graph but my graph consists of nbrVertex/2 components. So it is very sparse.

1- Does the random walk is suitable to such kind of graph?

What should i want to add to get the random walk be suitable to my graph? Does if add edges between very pairs of nodes with very small wait can help?

2- Suppose I added edges between nodes to make the graph strongly connected. On the latter, i execute the random walk, the difference between the old and the new vectors (p(t+1)-p(t)) still decreases (as it should be) till the iteration 10 and after they go to increases? DOES this means something? how can I interpret that?

user2869180
  • 129
  • 1
    I'm not sure what you mean by "can we use it" and "is it suitable". Of course you can run a random walk on a disconnected graph, but it will always stay in the same component where it started. – Nate Eldredge Jul 30 '19 at 04:58
  • @NateEldredge Can we use? you have already answer on that :). Suitable i mean is adequate for such kind of graph. – user2869180 Jul 30 '19 at 05:01
  • 1
    "Adequate" for what purpose? It depends on what you are trying to accomplish. – Nate Eldredge Jul 30 '19 at 05:02
  • @NateEldredge each node is composed of two component v=(v1,v2) where v2 is a potential candidate of v1. so im trying to get a rank for ech node using random walk – user2869180 Jul 30 '19 at 05:04

0 Answers0