Social Network

Analysis
AI 4144
Surabhi Agrawal
Assistant Professor
AIML Deparment

23/10/24 1
 SYLLABUS
1. Introduction to Social Web[1]: Nodes, Edges and Network measures, Describing Nodes and Edges, Describing
Networks, Layouts
2. Visualizing Network Features [1]: The role of Tie Strength, Measuring Tie Strength, Tie Strength and Network
Structure, Tie Strength and Network Propagation, Link Prediction, Entity Resolution
3. Link Prediction[1]: Case Study Friend Recommendation, Introduction to Community Discovery, Communities
in Context, Quality Functions;
4. Algorithms[2]: The Kernighan-Lin algorithm, Agglomerative Algorithms, Spectral Algorithms, Multi-level
Graph Partitioning, Markov Clustering, Other Approaches;
5. Introduction to Social Influence[2]: Influence Related Statistics, Social Similarity and Influence, Homophile,
Existential Test for Social Influence, Influence and Actions, Influence and Interaction, Influence Maximization in
Viral Marketing.
Reference Text Books
1. J. Goldbeck, Analyzing the Social Web, Morgan Kaufmann, 2013.
2. C. C. Aggarwal, Social Network Data Analytics, Springer, 2011.

23/10/24 2
Describing Networks
BASIC STRUCTURES AND PROPERTIES

23/10/24 3
Describing Networks : Basic Structures and Properties
• Building Blocks: Nodes, Edges
• Basic Structures
• Important
• Interesting
• Useful

23/10/24 4
Basic Structures: Subnetwork
• Subset of
• Nodes
• Edges

23/10/24 5
Subnetwork : Singleton
• Simplest form of subgraph/network
• Consists of a single node (Exactly one)
• Contains no edges
• Isolated from other nodes
• These represent people who signed up
for an account to access some part of
the site other than the social networking
features, or people who signed up but
never actively participated.

23/10/24 6
Subnetwork : Dyad
• Also known as two vertex graph
• Consists exactly two nodes
• Undirected
• zero edges (an isolated pair of vertices –
Null Dyad
• one edge (a connection between the two
vertices) – Connected Dyad
• Directed
• One edge in any direction – Asymmetric
Dyad
• Two edges in both directions – Mutual
Dyad

23/10/24 7
Subnetwork : Triad
• Consists exactly three nodes, subgraphs of
order three
• Undirected Triads
• Zero edges – Null Triad
• Only One edge (A pair and a stranger) –
disconnected Triad
• Two edges (A pair of strangers to eachother) –
Open Triad
• Three Edges (No Strangers) – Closed Triod

23/10/24 8
Subnetwork : Triad - Directed

23/10/24 9
Subnetwork : Clique
• A subset of vertices in a graph where every
pair of vertices are mutually connected
• For a graph or subgraph to be a clique, every node
must be connected to every other.

23/10/24 10
Subnetwork : Cluster
• A subset of vertices in a graph which are closely connected to each other.
• No standard/formal definition
• Shows presence of cohesive or tightly connected subgroups within a larger graph

23/10/24 11
Subnetwork : Egoistic Network
• Also referred to as “ego network”
• represents the connections (alters)
surrounding a single individual or entity in
a large network
• It has a depth
• Can be directed or undirected
• Degree-1 egocentric network-It only
shows us the nodes D is connected to.
• 1.5-degree egocentric network- If we want
to see only D’s neighbors and their
connections.
• 2-degree egocentric network- all of D’s
neighbors, their connections to one
another, and all of their neighbors.
23/10/24 12
Basic Structures: Other Concepts
• Path: A path in a graph G is a sequence of vertices v₁, v₂, v₃, ...,
vₖ, where each pair of consecutive vertices (vᵢ, vᵢ₊₁) is connected
by an edge in G.
• Number of edges in a path gives its length
• Simple path has no repeated vertices
• In a Cycle, path starts and ends at same vertex
• Connected Graph: In a connected graph, there exists at least
one path between any pair of distinct vertices.
• every vertex is reachable from every other vertex.
• Disconnected Graph: at least two vertices or groups of vertices
that have no path connecting them.
• One or more nodes are isolated from rest of the network

23/10/24 13
Bridges and Hubs
Bridges:
• bridge is an edge that connects two otherwise
separate groups of nodes in the network.
• edge between nodes P and F is a bridge because
if we take it out, the group of nodes on the right
will be totally disconnected from the group of
nodes on the left.
Hubs:
• Most connected nodes in the network.
• In Figure, node P would be a hub because it has
many connections to other nodes.

23/10/24 14
Network Measures
BASIC, DEGREE BASED, DIRECTED GRAPH, EGO NETWORK

23/10/24 15
Network Measures: Basic
• Graph Order: The order of a graph (typically written as n), is the number of nodes in
the graph
n=|V|
• Graph Size: The size of a graph (typically written as m), is the number of edges in the
graph
m=|E|
Maximum Size of an undirected graph is defined as follows

Maximum Size of a directed graph is defined as follows

23/10/24 16
Network Measures: Basic
• Graph Density: Represents the measure of original graph size compared to its
maximum possible size. Useful in identifying social differences among equally sized
communities.
Density in Undirected Graphs:

Consider the example graph with n = 9 and m = 16

Density in Directed Graphs:

23/10/24 17
Next Lecture
Detailed discussion on Network measures

23/10/24 18
 Thank You J …..

Any Questions?

23/10/24 19