Introduction - Social Networks
- Social Network
- Actors: Node, Point, Vertices
- Relational Ties: Links, Lines, Edges, Arcs
- Strength
- Information capacity
- Rate of flow
- Distances
- Probabilities
- Frequency
- Dyad: Pair of actors and the possible tie between them
- Triad: Subgraph consisting of three nodes and the possible lines among them
- A triad may be in one of four possible states
- Granovetter's model
- Forbidden triad: Triad with two lines present and one line absent
- Subgroup: Any subset of actors and all ties among them
- Group: Collection of all actors on which ties are to be measured
- Relation: How each actor is related to other actors on this relation
- Dichotomous relation: Either the relation among actors/nodes present or absent in a graph
- Directional ties: Relationship from actor A to B is distinctive from the relationship from actor B to A
- Social Network Data
- Types of Variables
- Structural: Measured on pairs of actors
- Composition: Measurements of actor attributes
- Mode: Distinct set of entities on which the structural variables are measured
- Affiliation Variables: The two modes are the actors and the events
- Boundary specification and sampling
- Realist approach
- Nominalist approach
- Snowball network sampling
- Network data, measurement and collection
- Issues in Measurement
- Unit of observation: actor/dyad/triad/subset of actors/network
- Modeling unit: actor/dyad/triad/subset of actors/network
- Relational quantification
- directional vs. non-directional/dichotomous vs. valued
- Data Collection
- Questionnaire
- roster vs. free recall
- free vs. fixed choice
- ratings vs. complete ranking
- Interview
- Observation
- Archival Records
- Special network designs
- Cognitive social structure
- Respondents give information on their perceptions of other actor’s network ties
- Experimental: Selected actors (and specified pairs)
- Ego-centered: Egos and alters
- Small world: Length of the chain and the characteristics of the actors
- Diary: Personal network
- Longitudinal Data Collection: How ties in a network change over time
- Measurement Accuracy, Validity, Reliability, Error
- Applications
- Web Graph
- Ranking a Node
- Google Page Ranking is based on this
- Routing in peer‐to‐peer networks
- Marketing and advertising
- Social Network
- Link Prediction
- Friend suggestions in Facebook
- Trust and distrust
- Diffusion of information and epidemics
- Networks
- Examples
- Ingredients Network
- Making edge between two ingredient which are part of a dish
- Community Structure
- Synonymy Network
- Making edge between two words which are synonyms
- Degradation of Synonymity
- Friendship Network
- Road Network
- Email Network
- Citation Network
- Collaboration Network
- Searching in Network
- Complexity is very high
- Considering the total number of graph possible between 50 nodes is more than the total number of atoms in the universe
- Small World Phenomenon
- Complexity => O(log2n)
- Network Schemas
- Graph Theoretic
- Sociometric
- A social network data set consisting of people and measured affective relations between people
- Sociometry: Study of positive and negative affective relations
- Sociomatrices: Relational data presented in two-way matrices
- Two dimensions are indexed by the sending actors (rows) and the receiving actors (columns)
- In Three dimensions layers index the relations
- Sociagram: Graphic representation of social links that a person has with others
- Algebraic
- Relations are represented with distinct capital letters
- Datasets
- Zachary Karate Club: Using just 34 nodes for friendship network
- Different Formats
- CSV, GML, Pajek Net, GraphML, GEXF