Social Networks (IT-703 (C)) - Unit 3 Short Notes

UNIT 3: SOCIAL NETWORKS - EXAM-FOCUSED SHORT NOTES

Based on RGPV Paper C: Social Networks (Nov 2023) Analysis

I. FOUNDATIONS & EVOLUTION OF SOCIAL NETWORKS

1.1 Emergence and Historical Development of the Social Web

• Definition: The Social Web refers to the evolution of the World Wide Web from a static information repository (Web 1.0) to a platform for user-generated content, interaction, and collaboration (Web 2.0), culminating in an intelligent, interconnected web of data (Web 3.0/Semantic Web).

• Key Evolutionary Phases:

  • Web 1.0 (Read-Only): Static HTML pages, limited user interaction (e.g., early company websites).

  • Web 2.0 (Read-Write): Rise of user-generated content, rich internet applications, and participatory culture. Enabled by APIs, AJAX, and tagging. Examples: Facebook, YouTube, Wikipedia, Blogger.

  • Web 3.0 (Read-Write-Execute): Focus on machine-readable data, semantic interoperability, and intelligent agents. Built on Semantic Web standards (RDF, OWL) to create a "web of data."

• > [!TIP]

  • Exam Focus: Be prepared to contrast Web 1.0 vs. 2.0 vs. 3.0. The Social Web is synonymous with Web 2.0's shift to user participation.

1.2 Classification and Types of Web-Based Networks

• Social Networks: Focus on relationships and interactions between individuals/groups (e.g., Facebook, LinkedIn). Structure is based on ties (friendship, professional).

• Information Networks: Focus on linking and organizing information objects. Structure is based on hyperlinks or citations (e.g., the World Wide Web itself, academic citation networks like Google Scholar).

• Knowledge Networks: Focus on connecting concepts, ideas, and expertise. Structure is based on semantic relationships (e.g., ontologies, expert networks, collaborative tagging systems like Delicious).

• > [!TIP]

  • Exam Focus: A single platform can embody multiple types. E.g., Wikipedia is an information network (articles link to each other) and a social network (editors collaborate).

1.3 Conceptual and Architectural Frameworks

• Conceptual Framework: Views a social network as a graph $$\displaystyle G = (V, E) $$, where $V$ are actors (nodes) and $E$ are relationships (edges). Analysis focuses on graph properties (density, diameter, clustering).

• Architectural Framework (Layered):

  1. Presentation Layer: User interfaces (web, mobile apps).

  2. Application Layer: Core services (profiles, messaging, news feeds).

  3. Data Layer: Storage for user data, relationships, content (often graph databases).

  4. Infrastructure Layer: Servers, networking, cloud platforms.

• > [!TIP]

  • Exam Focus: Know the graph representation $$\displaystyle G = (V, E) $$. For "conceptual framework," think graph theory.

II. SOCIAL NETWORK ANALYSIS (SNA) CORE CONCEPTS & METHODS

2.1 Importance, Objectives, and Applications

• Importance: Moves beyond descriptive statistics to uncover hidden structures, influential actors, sub-groups, and information flow patterns.

• Objectives: Identify key players, detect communities, understand resilience, model diffusion of information/behavior.

• Applications: Marketing (targeting influencers), cybersecurity (detecting botnets), public health (tracking disease spread), organizational management (mapping collaboration).

2.2 Centrality Measures: Types and Significance

Measures identify the most "important" nodes within a network.

• $n$: total nodes, $d(v,u)$: shortest path length, $$\displaystyle \sigma_{st} $$: total shortest paths from $s$ to $t$, $$\displaystyle \sigma_{st}(v) $$: paths through $v$, $A$: adjacency matrix, $\lambda$: eigenvalue.

• > [!TIP]

  • Exam Focus: Know the intuition behind each measure, not just formulas. Betweenness is for bridges, Eigenvector for influence. Be ready to compute simple examples.

2.3 Clustering & Community Detection

• Clustering Coefficient (Local): Measures the degree to which a node's neighbors are connected to each other.

$$C_i = \frac{2T_i}{k_i(k_i - 1)}$$

*   $$\displaystyle T_i $$: number of triangles through node $i$, $$\displaystyle k_i $$: degree of node $i$.

*   **Global Clustering Coefficient:** Average of all $$\displaystyle C_i $$.

• Community Detection: Finding groups of nodes with dense internal connections and sparser external connections.

  • Modularity ($Q$): Key metric to evaluate community structure quality.

$$Q = \frac{1}{2m} \sum_{ij} \left[ A_{ij} - \frac{k_i k_j}{2m} \right] \delta(c_i, c_j)$$

    *   $$\displaystyle A_{ij} $$: adjacency, $$\displaystyle k_i, k_j $$: degrees, $m$: total edges, $\delta$: 1 if nodes $i,j$ in same community.

    *   **Interpretation:** $Q$ close to 1 indicates strong community structure.

• > [!TIP]

  • Exam Focus: Distinguish clustering coefficient (local density around a node) from modularity (global quality of a community partition).

2.4 Graph/Matrix Representation

• Adjacency Matrix ($A$): $n \times n$ matrix where $$\displaystyle A_{ij} = 1 $$ if edge $(i,j)$ exists, else 0. For undirected graphs, $A$ is symmetric.

  
  A = \begin{bmatrix}
  
  0 & 1 & 0 \\
  
  1 & 0 & 1 \\
  
  0 & 1 & 0
  
  \end{bmatrix}
  
  

• Incidence Matrix ($B$): $n \times m$ matrix ($m$ = edges). $$\displaystyle B_{ij} = 1 $$ if node $i$ is incident to edge $j$, else 0.

• Use: Matrix operations (e.g., $$\displaystyle A^k $$ gives number of paths of length $k$ between nodes) enable computational analysis.

• > [!TIP]

  • Exam Focus: Be able to draw adjacency matrix for a given small graph. Know that $$\displaystyle A^2 $$ counts 2-step walks.

III. SEMANTIC WEB & ONTOLOGIES FOR SOCIAL DATA

3.1 Resource Description Framework (RDF) & RDF Schema (RDFS)

• RDF: Foundation of Semantic Web. Represents data as triples: (Subject, Predicate, Object).

  • Example: (Alice, knows, Bob).

  • Serialized in Turtle, RDF/XML, N-Triples.

• RDFS: Extends RDF to define vocabularies (schemas).

  • Key Elements: rdfs:Class (defines a type), rdfs:subClassOf (hierarchy), rdfs:domain/rdfs:range (constraints on predicates).

  • Limitation: Limited expressiveness (no cardinality, disjointness).

• > [!TIP]

  • Exam Focus: RDF is data model (triples), RDFS is schema language. Know the triple structure.

3.2 Web Ontology Language (OWL)

• Unique Features & Capabilities: More expressive than RDFS. Based on Description Logics.

  • Key Features:

    • Class Equivalence & Disjointness: owl:equivalentClass, owl:disjointWith.

    • Property Characteristics: owl:TransitiveProperty, owl:FunctionalProperty, owl:inverseOf.

    • Cardinality Restrictions: owl:minCardinality, owl:maxCardinality.

    • Complex Class Expressions: Intersection (owl:intersectionOf), Union, Complement.

  • Use: Formal, machine-interpretable knowledge representation for complex domains (e.g., medicine, advanced social profiles).

• > [!TIP]

  • Exam Focus: OWL adds logical constraints and richer relationships missing in RDFS. Mention Description Logics as its foundation.

3.3 FOAF (Friend of a Friend)

• Foundation: A popular RDF-based ontology for describing people, their activities, and relationships.

• Key Classes & Properties:

  • foaf:Person (or foaf:Agent).

  • foaf:name, foaf:mbox (email), foaf:homepage.

  • foaf:knows (social link), foaf:member (group membership).

  • foaf:Image (depiction), foaf:Project (involved in).

• Significance: Enables decentralized social networking and data portability. A FOAF profile is a machine-readable "business card" that can be linked across the web.

• > [!TIP]

  • Exam Focus: FOAF is the practical application of Semantic Web for social data. It uses RDF triples to represent social graphs.

IV. COMMUNITY DETECTION, EVOLUTION & NETWORK TRANSFORMATION

4.1 Definitions of "Community"

• Local (Structural): A set of nodes with dense internal connections (high clustering). Detected via modularity optimization.

• Global (Functional): A set of nodes sharing a common property or role (e.g., all students in a department). Requires external attribute data.

• Vertex-Centric: A community is the set of nodes reachable from a given seed node within a certain distance or conductance threshold. Basis for many algorithms (e.g., label propagation).

• > [!TIP]

  • Exam Focus: The question specifically asks for these three perspectives. Memorize: Local=Structure, Global=Attribute, Vertex-Centric=From a seed node.

4.2 Extracting Community Evolution (from Web Archives)

• Goal: Track how communities form, grow, merge, split, and dissolve over time.

• Key Metrics & Methods:

  1. Temporal Slicing: Treat each time snapshot (archive) as a static network $$\displaystyle G_t $$.

  2. Community Tracking: Match communities across $$\displaystyle G_t $$ and $$\displaystyle G_{t+1} $$ using Jaccard similarity of node sets.

  3. Evolution Metrics:

     • Size: Node count change.

     • Density: Internal edge density.

     • Lifespan: Duration of existence.

     • Stability: How member composition changes.

     • Event Types: Birth, Death, Merge, Split, Growth, Contraction.

• > [!TIP]

  • Exam Focus: Focus on the process: snapshot → detect communities → track across time → measure change. Mention Jaccard index for matching.

4.3 Network Reduction Techniques

• Purpose: Simplify large, dense networks to reveal core structure, reduce computational cost, and improve visualization.

• Key Methods:

  • k-Core Decomposition: Recursively removes nodes with degree $$\displaystyle < k $$. The k-core is the maximal subgraph where all nodes have degree $\geq k$. Reveals the dense core.

  • Backbone Extraction: Removes "noisy" or statistically insignificant edges based on weight (e.g., using Disparity Filter for weighted networks).

  • Bipartite Projection: Projects a bipartite network (e.g., users-items) onto one set (e.g., user-user network based on shared items).

• > [!TIP]

  • Exam Focus: k-Core is the most important reduction technique for social networks. It finds the "heart" of the network.

V. APPLICATIONS & ENABLING HUMAN-CENTRIC EXPERIENCES

5.1 Role of Social Networks in Human Experiences

• Facilitates maintaining relationships (weak ties), identity expression, collective action, and access to diverse information.

• Enables experiences like real-time coordination, crowdsourcing, and virtual communities of interest.

5.2 Reality Mining

• Concept: Collection and analysis of real-world, sensor-based data (from mobile phones, wearables) to understand human behavior, social interactions, and mobility patterns.

• Techniques: Call detail records, Bluetooth proximity, GPS location, accelerometer data.

• Applications: Urban planning, epidemic modeling, targeted advertising, productivity analysis.

• > [!TIP]

  • Exam Focus: Reality Mining = sensing + mining of real-life behavior (not just online activity).

5.3 Context Awareness

• Importance: Systems that adapt behavior based on user's context (location, time, activity, social setting, device).

• Implementation in Social/Mobile Networking:

  • Location-based services (check-ins, nearby friends).

  • Adaptive content delivery (show different feeds at work vs. home).

  • Privacy controls (automatically hide location from certain contacts).

  • Requires sensors (GPS, accelerometer), context models, and reasoning engines.

• > [!TIP]

  • Exam Focus: Context Awareness = "who, what, where, when, why" of the user. It's the bridge between physical world and social apps.

VI. SECURITY, PRIVACY & ATTACK SPECTRUM IN ONLINE SOCIAL NETWORKS (OSNs)

6.1 Privacy in OSNs: Issues & Challenges

• Major Issues:

  • Data Collection & Profiling: Extensive personal data harvesting.

  • Disclosure Risks: Accidental sharing (e.g., geotagged photos), de-anonymization.

  • Complex Privacy Settings: Users often confused or unaware.

  • Third-Party Apps: Access to profile/friends data.

  • Cross-Platform Correlation: Aggregating data from multiple sites.

• Challenges: Balancing utility vs. privacy, user education, designing usable privacy controls, regulatory compliance (GDPR).

• > [!TIP]

  • Exam Focus: Link privacy issues to data types (profile, connections, content, metadata).

6.2 Attack Spectrum & Countermeasures

• > [!TIP]

  • Exam Focus: Cloning = fake copy of existing profile. Hijacking = take over real account. Porting = moving profile across platforms.

6.3 Censorship Attacks

• Types:

  • Content Filtering/Blocking: Preventing specific posts, pages, or entire sites from being accessed.

  • Connection Tampering: Resetting connections, throttling bandwidth to OSNs.

  • Account Suspension/Deletion: Forcibly removing user accounts or content.

  • Search Result Manipulation: De-listing or demoting content.

• Impact: Suppresses free speech, limits information flow, fragments public discourse.

• > [!TIP]

  • Exam Focus: Censorship attacks are typically state-sponsored or platform-enforced actions to block access or remove content.

6.4 Challenges for Decentralized OSNs (DOSNs)

• Technical: Data synchronization, identity management without central authority, scalability, spam/abuse moderation in a distributed system.

• Social: User adoption (network effects), critical mass problem, user experience consistency, trust establishment between nodes.

• Architectural: Designing protocols for privacy (e.g., end-to-end encryption for posts), content discovery without central index, economic models for hosting.

• > [!TIP]

  • Exam Focus: The core challenge is the "chicken-and-egg" problem of network effects—users won't join unless their friends are there, but friends won't join unless users are there.

VII. SUPPORTING TECHNOLOGIES & PROTOCOLS

7.1 Email Groups (Mailing Lists)

• Structure: One email address distributes messages to a list of subscribers.

• Functionality: Asynchronous, threaded discussions. Moderated or unmoderated. Archives are searchable.

• Role as Social Tool: Creates persistent, topic-focused communities. Pre-dates modern forums. Examples: listserv, Google Groups.

• > [!TIP]

  • Exam Focus: Email groups are early, text-based, asynchronous social networking tools focused on topic-based discussion.

7.2 RSS Feeds

• Concept: Really Simple Syndication (or Rich Site Summary). A web feed format for publishing frequently updated content (blogs, news, podcasts).

• Mechanism: Publisher creates an XML file (RSS feed) with metadata (title, link, description, date). Users subscribe via an RSS reader/aggregator, which checks for updates.

• Use in Networks: Enables content aggregation from multiple sources. Users can follow many blogs/news sites in one place, creating a personalized information network. Powers podcast distribution.

• > [!TIP]

  • Exam Focus: RSS is a pull-based (user's reader polls) content distribution mechanism, not a social interaction platform itself.

\boxed{\text{End of Unit 3 Notes - Focus on Past Paper Questions C (Nov 2023)}}