Statistical Mechanics

of Complex Systems
Dr. Suweis Samir
Prof. Amos Maritan

samir.suweis@unipd.it - office 329: Friday 9.30-10.30

amos.maritan@pd.infn.it - office 326: Thursday 14.30-15.30

 What is a Network?
Network (graph) is a mathematical structure
composed of points connected by lines

Network Theory <-> Graph Theory

Network ↔ Graph

Nodes ↔ Vertices (points)

Links ↔ Edges (Lines)

System vs. Parts = Networks vs. Nodes

Real networks can be divided in four different classes:

Describing a network formally!
N nodes and E edges,!

where E ≤ N(N-1)/2 !

N = 7, E = 9!

Note: In graph theory language this graph is of order 7 and size 9.!
 Adjacency matrix!
The most convenient way of describing a
network is the adjacency matrix aij. !

A link between node i to node j is recorded by a

‘1’ in the ith row and the jth column.!
 Adjacency matrix!
Undirected networks
have a symmetric
adjacency matrix !
aij = aji.!

Directed networks in
general have
asymmetric aij.!
 Self-interactions!

Networks also can have self-interactions, which

correspond to the diagonal entries aii.!

If we allow self-interactions, we can have up to

E = N2 edges.!
 Weighted networks!
In a weighted network a real number is attached to each
edge, so that we obtain a real adjacency matrix, usually
denoted as wij.!
 Networks As Graphs
Type of graphs/networks
Some Details

Networks As Graph
• A-cyclic networks. Directed
Acyclic Directed Graph
1
2

5 4 3

Some Undirected
Bipartite Graphs Details (trees)

2
Path: The length of a path is the number of edges
1
traversed in the
path i->j->k —>
Tree Aij Ajk=1
Graphs
Loops: path closing in the starting node.
Acyclic Directed Graph
5 4 3
Planar Networks
• Bipartite graph
Bipartite
Incidence Graphs
matrix B and projection graphs O
Bij -> Oij(1), Oij(2)
Technological Networks
Transporta3on networks: airlines
Network of actor co-starring in movies
 Information Networks
Paper Cita'ons Networks
Social
Networks
 Internet as measured by Hal Burch and Bill Cheswick's Internet Mapping Project.

5,5*109
nodes
(indexed)
 Biological Networks

C. Elegans Gene-Protein Network

Freshwater food web
The human brain network
1011 neurons!

<103> dendrites!
 Properties of real world networks
• SMALL WORLD: small average path length.
Milgram Experiment (1967)
Co-authorship studies [ERDOS number]
WWW - average path length 3.1 (50 M pages)
Facebook - mean 4.74 (721 million users) - Bockstrom et al 2012
• SCALE FREE DISTRIBUTION: existence of hubs.
Power law degree distribution (but finite size effect)
• TRANSITIVITY: high in social networks
Average probability that two neighbours of a vertex are
themselves neighbours.
Reciprocity: frequency of loops of length 2
Assortative/Dissasortative mixing. Correlations among nodes
• MODULAR STRUCTURE: Existence of communities
LOCAL AND GLOBAL GRAPH
TOPOLOGICAL PROPERTIES
 Degree!
In an undirected network the degree ki of a node
i is the number of nodes i is connected to:!
ki = Σj aij = Σj aji!

Here k1 = 2, k2 = 4, k3 = 1, k4 = 3 and k5 = 2.!

 Connectivity Descriptors
Local (node) centrality: vertex (node) degrees, ki

n
ki = ∑ aij = ∑ j − neighbors aij
j =1

Global (Network) descriptors: number of links, m

€
n n n
1 1
L(G) = ∑ k i = ∑ ∑ aij
2 i=1 2 i=1 j =1

€
Connectivity Descriptors-2
Average and Normalized Descriptors

Average vertex (node) degree:

< ki >= c =2L /N

Network connectance (density)

L 2L k
€ C(G) = N = =
( 2 ) N(N −1) (N −1)
 Strength!
In a weighted, undirected network the strength is the
sum of the weights for the edges connecting to a node:!
si = Σj wij = Σj wji!

Hence s1 = 4, s2 = 18, s3 = 2, s4 = 13 and s5 = 15.!

 In-degree and out-degree!
In a directed network the in-degree ki(in) of a
node i is the number of directed edges pointing
to node i:!
ki(in) = Σj aji!

while the out-degree ki(out) of a node i is the

number of directed edges pointing from node i:!

ki(out) = Σj aij!
 In-degree and out-degree!
Thus, in a directed network, nodes can be
highly connected, yet also isolated (e.g. in terms
of sending or receiving information.) !
 Distance!
The distance between two nodes i and j is the shortest
path connecting the two nodes.!

dij = 4!
 Diameter!
The diameter of a network is the largest distance in the network - in
other words it is the maximum shortest path connecting any two
nodes.!

D=2 D = 1!

Note: Fully connected networks (like the one on the right) have
diameter D = 1.!
 Distance Descriptors
Node descriptors:
N

node distance, di
di = ∑ dij
j=1

Node eccentricity, eei ei = max j∈N (dij )

Network descriptors:
N N N

Network distance, D(G) D(G) = ∑ di = ∑∑ dij

i=1 i=1 j=1

Network diameter, Diam(G) Diam(G) = Maxij∈N (dij )

Network radius, Rad(G) Rad (G ) = Min (ei ) = Min ( Max(d ij ))

 Distance Descriptors-2
Average and Normalized Descriptors

Average node distance, <di>

< di > = D(G) /N

Average network distance, <d>

(average degree of separation, average path length)

€
< d > = D /N(N −1)

€
 Distances in Directed
Networks
!  In-distances and out-distances

0,3 D(in) = D(out) = 6

1
<d> = 6/(4x3) = 0.5 ???
2 -2,1
d21=? d13=?
3 4 0, 2 -4, 0

!  Some distances in (directed) graphs are equal to infinity !

!  Vertex Components: Weakly connected, Strongly Connected,

 1
G
What is the connectance of G?
2 3 4 5

A C=0.4 B C=4/5 C C=4/7

What is minimum node distance?

A di=1 B di=5 C di=8

What is average path length?

A <d>=1.8 B <d>=7.2 C <d>=8/5

 1 L= 4

<k>= 8/5 = 1.6

C –connectivity
1 3 2 1
C = 1.6/4 = 0.4 = 40%

A C=0.4 B C=4/5 C C=4/7

1 1 2 3 4 5 di
1 0 2 1 2 3 8
D(G) =
2 2 0 1 2 3 8 di – node distance
3 1 1 0 1 2 5
2 3 4 5 4 2 2 1 0 1 6 (node distance degree)
5 3 3 2 1 0 9

A di=1 B di=5 C di=8

8 D = 8+8+5+6+9 = 36
<d>– average distance degree
<di> = 36/5 = 7.2
8 5 6 9
<d > = 36/5.4 = 1.8

A <d>=1.8 B <d>=7.2 C <d>=8/5

What is the diameter of a ring of nodes?

A D(G)=1

B D(G)=N

D(G)=N/2 or N/2-1
C depending on N
What is the diameter of a ring of nodes?

A D(G)=1

B D(G)=N-1

D(G)=N/2 or N/2-1
C depending on N
 What is the diameter of a tree?

A D(G)=K

B D(G)=KLog[N]

C D(G)=2log2(N+1)-2

Tree with N nodes

and K levels D D(G)=2(N-1)K
 What is the diameter of a tree?

A D(G)=K

B D(G)=KLog[N]

diameter is 2K C D(G)=2log2(N+1)-2

K levels has N = eK+1-1

nodes so, K = log2(N+1) -1 D D(G)=2(N-1)K
 Clustering coefficient!
The clustering coefficient measures how densely connected the
neighbourhood of a node is. !

It does this by counting the number of triangles of which a given node i is

a part of, and dividing this value by the number of edge pairs. !

ci = [2/ki (ki - 1)] Σjk aij ajk aik!

Often the clustering coefficient is averaged over the entire network:!

C = (1/N) Σijk [2/ki (ki - 1)] aij ajk aik!

Where N is the number of nodes.!

 Betweenness!
The rather awkward word betweenness is a measure of the
importance of a node or edge. !

The most widely used is shortest-path betweenness:!

Consider a pair of nodes, and all shortest paths between them.
For any given edge or node in the network, we can determine
the fraction of these shortest paths which pass through it. The
shortest-path-betweenness is the sum of these fractions over
all pairs. !

Other forms include random-walk betweenness and current-flow

betweenness.!
 Betweenness: an example!
While betweenness of a given node or edge is calculated
over all pairs of nodes, consider the contribution
associated with one particular node (s below):!

(a)  In a tree, the betweenness !

is rather straightforward.!

(b)  In a network with loops,!

the betweenness becomes!
more complicated, e.g.!
25/6 = 1 + 1 + 1 + 1/2 + 1/3 + 1/3 !

Image: Newman and Girvan, PRE 69, 026113 (2004)

 Nodes Correlation
•  Correlations is a “similarity” measure between two vertices

•  Pearson Correlation: # of common neighbors vertices have

minus the expected number that they would have if they chose
their neighbors at random

• Networks in which highly connected nodes are linked to

other nodes with a high degree are termed assortative (social)

• Networks in which highly connected nodes are only linked to

nodes with a low degree are termed disassortative
(internet and biological)
 Nearest-neighbour degree!
The nearest neighbour degree knn of a node i is the average
degree of the neighbours of i. !

The average nearest neighbour degree 〈knn〉 is knn averaged

over all nodes of the same degree k.!

Assortativity can also be measured by plotting the average

nearest neighbour degree 〈knn〉 as a function of the degree k.!

An increasing slope indicates assortativity while a

decreasing one signals disassortativity.!
 kj , are connected, can be calculated without probl
two stubs (i.e. half-edges), incident with i and j. T
Modularity
there are ki stubs incident with i out of a total of 2
product pi pj , since edges are placed independently
Pij = 2mpi pj = ki kj /2m of edges between i and j. S
X ✓ ◆
1 ki kj
Q = Aij (Ci , Cj ).
2m ij
2m

6 The difference is that the configuration model maintains the

Physics Reports 486 (2010) 75–174
of modularity the degree sequence of a realization is in general
at hand. The
Contents lists twoatmodels
available ScienceDirectare equivalent in the limit of infinite g

Physics Reports
journal homepage: www.elsevier.com/locate/physrep

Community detection in graphs

Santo Fortunato ⇤
Complex Networks and Systems Lagrange Laboratory, ISI Foundation, Viale S. Severo 65, 10133, Torino, I, Italy

article info abstract