Lecture3

Network Science
Historical Background and network analysis metrics
“Measure what is measurable, and make measurable what is not so”

Galileo Galilei
Graph theory:
Birth of Network Science
Euler, is known as the father of graph theory: 1741
 The problem of seven bridges of Konïgsberg (his home)

2
Birth of Network Science (Cont..)
 A solution path cannot exist on a graph that has more than two nodes with an
odd number of links
 Euler declared: “The existence of the path does not depend on our cleverness to
find it. Rather it is a property of the graph”
 Graphs (or networks) have properties hidden in their construction that limit or
enhance our ability to do things with them
 In 19th century, the advancement in the research field of graph theory

continued by many mathematician including William Hamilton, Gustav
Kirchoff, Thomas kirkman and Aruther Cayley (Ghoshal, 2009; Newman,
2011)
3
 Their research and development in graph theory was based on theoretical
mathematical objects as compared to the problems related to real-world
systems
 Historically, the social scientists realized very earlier in the beginning of

20th century, the applications of network analysis and used sociograms as
networks
 The first person who as a sociologist and psychiatrist carried the

systematic study of social relations as network was Jacob Moreno
 He systematically collected and analyzed the pattern of interactions
in very small groups like school class rooms and worker groups
(Moreno, 1953)
 By observing the structure of social interactions and by drawing on

the paper as nodes and links one can predict the future behavior or
functioning of the network
 He called his networks as sociograms in which nodes were

represented as people and links were created if they had some kind
of acquaintances
Sociogram made by Jaccob Moreno
Milgrams Six degrees sepration
 In 1967, Stanley Milgram, a social psychologist at Harvard,
was interested in learning how many acquaintances it would
take to connect two randomly selected individuals from the
population
 A group of people from Omaha (Nebraska) and Wichita

(Kansas) was asked to send a letter to an unknown person in
Boston (Massachussetts)
 Basic Rule of the experiment:
 People should forward the letter to a person that they consider closer to the target person
 Results of one experiment
 232 out of 296 letters never reached the target
 64 letters reached the target (with paths from 2 to10)
 The average path length was….5.2 (steps)

This is big world This is small world
8
 The quantitative analysis and research of social scientists helped in further development
of mathematical techniques
 Anatol Rapoport, a mathematical psychologist argued that the frequency distribution of

links related with nodes in the network plays very important role to measure properties of
network
 With his collaborator R. Solomonoff he proposed random graph which was a simple
model of network (Solomonoff and Rapoport, 1951)
 His model then independently studied by Paul Erdos with his collaborator Alfred Renyi
(Erdo¨s and R´enyi, 1960)
 The ER model remained very popular from late Fifties to late Nineties and, the
modeling of almost all complex networks was based on random graph theory
9
Network Analysis Metrics
Node degree and strength (Local measures) in social networks
The node degree is the number of links directly connecting to the node with other nodes,
in the network
 The degree ki of the node i can be computed as given in equation
where ki represents the degree of node i and a is the adjacency matrix and the entry aij
represent connectivity if the value is 1, otherwise it is 0 and N is the total number of nodes
in the network
10
moviegalaxies.com
11
12
trade in petroleum
and petroleum
products, 1998,
source: NBER-Unite
Nations Trade Data
Quiz Q:
• Which countries have high indegree (import
petroleum and petroleum products from many
others)
—Saudi Arabia
—Japan
—Iraq
—USA
—Venezuela
different notions of centrality
In each of the following networks, X has higher

centrality than Y according to a particular measure
indegree outdegree betweenness closeness

Node degree and strength (Local measures)
Weighted degree or strength of node is used (Yang and Knoke, 2001), (Barrat et al.,
2004a) in weighted networks, which can be defined as the sum of a node direct links weight
as given in equation
where w represents the weighted adjacency matrix. The value of the entry wij in matrix is
> 0, if the node i is connected with j, whereas the value show the weight of link
16
Node degree centrality in social networks:
The identification of main or critical nodes in complex networks analysis has remained
very important topic. Because, if a node is more central as compared to others in a network
can have three main advantages
1. Placement or location can potentially control the traffic/flow among other nodes
2. It can have more direct links
3. From that node, one can reach to all other nodes very quickly
 Further, as node strength do not consider the number of links with particular node,
therefore it gives only rough idea for the nodes actual contribution and involvement in
whole network from degree centrality point of view
17
 (Opsahl et al., 2010) has proposed a very realistic approach for degree centrality in
weighted networks by considering both the number of links and the total weight of links
(strength)
Where in this equation works as tuning parameter.
 If the value of alpha is 0 then it will only focus on the number of links connected with
nodes, and if the value is 1, then it will include the weight of links only
 If the value of alpha is in the range of 0 and 1, then it will measure both the weights and
degree, favorably
18

Lecture3

Uploaded by

Copyright:

Available Formats

Lecture3

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lecture3

Uploaded by

Copyright:

Available Formats

Network Science

Historical Background and network analysis metrics

“Measure what is measurable, and make measurable what is not so”

Euler, is known as the father of graph theory: 1741

 The problem of seven bridges of Konïgsberg (his home)

 In 19th century, the advancement in the research field of graph theory

 Historically, the social scientists realized very earlier in the beginning of

 The first person who as a sociologist and psychiatrist carried the

 By observing the structure of social interactions and by drawing on

 He called his networks as sociograms in which nodes were

 A group of people from Omaha (Nebraska) and Wichita

 Results of one experiment

 232 out of 296 letters never reached the target

 64 letters reached the target (with paths from 2 to10)

 The average path length was….5.2 (steps)

 Anatol Rapoport, a mathematical psychologist argued that the frequency distribution of

In each of the following networks, X has higher

indegree outdegree betweenness closeness

Where in this equation works as tuning parameter.

You might also like