Description
1. Random walk on random networks
(a) Create undirected random networks with 1000 nodes, and
the probability p for drawing an edge between any pair of
nodes equal to 0.01.
(b) Let a random walker start from a randomly selected node (no
damping). We use t to denote the number of steps that the
walker has taken. Measure the average distance hs(t)i of the
walker from his starting point at step t. Also measure the
standard deviation σ
2
(t) = h(s(t) − hs(t)i)
2
i of this distance.
Plot hs(t)i v.s. t and σ
2
(t) v.s. t. Here, the average h·i is over
all possible starting nodes and different runs of the random
walk (or different walkers). You can measure the distance of
two nodes by finding the shortest path between them.
1
(c) We know that a random walker in d dimensional has average
(signed) distance hs(t)i = 0 and q
hs(t)
2i = σ ∝
√
t. Compare
this with the result on a random network. Do they have similar relations? Qualitatively explain why.
(d) Repeat (b) for undirected random networks with 100 and
10000 nodes. Compare the results and explain qualitatively.
Does the diameter of the network play a role?
(e) Measure the degree distribution of the nodes reached at the
end of the random walk on the 1000-node random network.
How does it compare with the degree distribution of graph?
2. Random walk on networks with fat-tailed degree distribution
(a) Use barabasi.game to generate a network with 1000 nodes
and degree distribution proportional to x
−3
.
(b) Let a random walker start from a randomly selected node.
Measure and plot hs(t)i v.s. t and σ
2
(t) v.s. t.
(c) Are these results similar to results of random walks in d dimensional space? Explain why.
(d) Repeat (b) for fat-tailed networks with 100 and 10000 nodes.
Compare the results and explain qualitatively. Does the diameter of the network play a role?
(e) Measure the degree distribution of the nodes reached at the
end of the random walk on the 1000-node fat-tailed network.
How does it compare with the degree distribution of the graph?
3. PageRank
The PageRank algorithm, as used by the Google search engine,
exploits the linkage structure of the web to compute global “importance” scores that can be used to influence the ranking of search
results. Here, we use the random walk to simulate PageRank.
(a) For random walks on the network created in 1(a), measure
the probability that the walker visits each node. Is this probability related to the degree of the nodes?
2
(b) Create a directed random network with 1000 nodes, where
the probability p for drawing an edge between any pair of
nodes is 0.01. Measure the probability that the walker visits
each node. Is this probability related to the degree of the
nodes?
(c) In all previous questions, we used a damping parameter equal
to d = 1, which means no teleportation (because teleportation
probability is equal to 1 − d = 0). Now, we use a damping parameter d = 0.85. For random walks on the network created
in 1(a), measure the probability that the walker visits each
node. Is this probability related to the degree of the node?
4. Personalized PageRank
While the use of PageRank has proven very effective, the web’s
rapid growth in size and diversity drives an increasing demand
for greater flexibility in ranking. Ideally, each user should be able
to define their own notion of importance for each individual query.
(a) Create a directed random network with 1000 nodes, where
the probability p for drawing an edge between any pair of
nodes is 0.01. Use the random walk with damping parameter
0.85 to simulate the PageRank of the nodes.
(b) Suppose you have your own notion of importance. Your interest to a node is proportional to the the node’s PageRank,
because you totally rely upon Google to decide which website to visit (assume that these nodes represent websites).
Again use the random walk to simulate this personalized
PageRank. Here the teleportation probability to each node
is proportional to its PageRank (As opposed to the regular
PageRank, where teleportation probability to all nodes are
the same and equal to 1
N
). The damping parameter is equal
to d = 0.85. Compare the results with (a).
(c) More or less, (b) is what happens in the real world. However,
this is against the original assumption of normal PageRank,
where we assume that people’s interest in all nodes are the
same. Can you take into account the effect of this self-enforcement
and adjust the PageRank equation?
3
Useful commands:
(1) random.graph.game, netrw, get.shortest.paths, diameter; (2) barabasi.game, netrw,
get.shortest.paths, diameter; (3,4) random.graph.game, netrw, degree
If you are unsure, please just use the default values for the command parameters.
You can use the netrw package to implement random walks and personalized PageRank. If you
use igraph’s page.rank function for regular PageRank, please make sure you are using igraph version
0.6 or higher as it fixes a bug in the PageRank command. The netrw package source code for different
platforms are uploaded in the handouts section on course website. You can search online on how to install
R packages from source code for your specific operating system. If you are using RStudio you can just
install the package from archive in the package manager.
4
