Unit 5
🎯 Unit 5 Overview
Unit 5 focuses on social network graph mining. In this unit, we study how social networks
are represented as graphs and how useful patterns, communities and recommendations are discovered.
Exam Tip: Social network graph, community discovery, clustering and recommender system are very important for 7 and 14 marks questions.
📘 Social Network Graph Mining
Social network graph mining is the process of analyzing social network data using graph theory
and data mining techniques. It helps in finding relationships, groups, influential users and user behavior.
Simple Meaning: Social media users aur unke connections ko graph ke form me study karna hi social network graph mining hai.
🕸️ Social Network as Graph
A social network can be represented as a graph where users are represented as nodes and their relationships are represented as edges.
| Graph Term |
Social Network Meaning |
| Node / Vertex |
User, person, page or organization |
| Edge / Link |
Friendship, follow, like, comment or connection |
| Degree |
Number of connections of a user |
| Path |
Connection route between two users |
| Community |
Group of closely connected users |
⭐ Applications of Social Network Mining
- Friend recommendation
- Product recommendation
- Community detection
- Influencer identification
- Fraud detection
- Marketing and advertising
- Customer behavior analysis
- Fake account detection
- Sentiment analysis
- Information spread analysis
📂 Types of Social Networks
| Type |
Description |
Example |
| Friendship Network |
Users are connected through friendship. |
Facebook |
| Follower Network |
Users follow other users or pages. |
Instagram, X |
| Professional Network |
Users connect for jobs and professional growth. |
LinkedIn |
| Content Sharing Network |
Users share videos, images or posts. |
YouTube, Pinterest |
| Communication Network |
Users communicate using messages or calls. |
WhatsApp, Telegram |
🔗 Clustering in Social Network
Clustering means dividing users into groups based on similarity or connection strength.
Users inside one cluster are more strongly connected with each other.
Purpose of Clustering
- To identify similar users.
- To detect user groups.
- To improve recommendation systems.
- To understand user behavior.
- To support targeted marketing.
👥 Community Discovery
Community discovery is the process of finding groups of users that are highly connected
inside the group and less connected with outside users.
Example
In a college social network, students of the same branch or same class may form one community.
Importance
- Helps in finding social groups.
- Useful in recommendation systems.
- Helps detect fake or suspicious groups.
- Improves content targeting.
- Helps understand network structure.
📊 Difference Between Clustering and Community Discovery
| Clustering |
Community Discovery |
| Groups similar users or data points. |
Finds closely connected groups in a network. |
| Based on similarity measures. |
Based on graph connections. |
| Used in general data mining. |
Mostly used in graph and social network mining. |
| Example: grouping customers by buying behavior. |
Example: finding friend circles in Facebook. |
💡 Recommender System
A recommender system suggests items, products, friends, videos or services to users
based on their interests, behavior and relationships.
Examples
- YouTube video recommendations
- Amazon product recommendations
- Netflix movie recommendations
- Instagram friend suggestions
- LinkedIn job recommendations
⚙️ Types of Recommender Systems
| Type |
Description |
| Content-Based Filtering |
Recommends items similar to items liked by the user earlier. |
| Collaborative Filtering |
Recommends items based on behavior of similar users. |
| Hybrid Recommender |
Combines content-based and collaborative filtering. |
Important: Collaborative filtering is very commonly used in modern recommender systems.
✅ Advantages of Social Network Mining
- Helps understand user behavior.
- Improves recommendations.
- Supports business decisions.
- Helps identify influencers.
- Useful for fraud and spam detection.
- Improves targeted advertising.
⚠️ Challenges in Social Network Mining
- Huge amount of data
- Privacy issues
- Fake accounts and bots
- Dynamic nature of networks
- Noisy and incomplete data
- Real-time processing difficulty
- Security concerns
⭐ Important Questions
- Explain social network graph mining.
- How can a social network be represented as a graph?
- Write applications of social network mining.
- Explain types of social networks.
- What is clustering in social networks?
- Explain community discovery with example.
- Differentiate between clustering and community discovery.
- What is recommender system? Explain its types.
- Write advantages of social network mining.
- Write challenges in social network mining.
🔥 Last Minute Revision
- Social network graph: users = nodes, connections = edges.
- Degree means number of connections.
- Community means closely connected group.
- Clustering groups similar users.
- Community discovery finds strongly connected groups.
- Recommender system suggests items or users.
- Main recommendation types: Content-based, Collaborative, Hybrid.