Measures - Social Networks

  • Prominence: Reflects its greater visibility to the other network actors
    • Prestige
    • Centrality
      • Central actors
      • Peripheral actor
      • Centrality Measures
        • Degree: Nodes with higher value are more central in undirected graphs
          • Types
            • Indegree: Nodes with higher value are more central in directed graphs
            • Outdegree: Nodes with higher value are more prestigious in directed graphs
          • Measures
            • Actor level
            • Group degree: Quantifies the dispersion or variation among individual centralities
              • Freeman Group Degree Centrality
            • Variance of the degrees
            • Mean Degree
            • Standardized Average Degree = Mean Degree/(g-1) = Density
        • Closeness
          • Fairness/Peripherality: Total distance that actor i is from all other actors
          • Actor Closeness = 1/Fairness
          • Group Closeness
            • Freeman's general group closeness index
          • Improved actor-level centrality closeness index
        • Betweenness
          • Group Betweenness
            • Freeman's group betweenness centralization index
        • Eigen Vector Centrality (Eigencentrality) Measure: Measure of the influence of a node in a network
          • Steps
            • Construct matrix A, representing connection between nodes based on given graph
            • Solve for |A - λI| = 0, to get non-zero eigen values
            • Solve corresponding to largest (Principal) eigen value, (A - λI)V = 0
            • V is the eigen vector, Assume one value to be "t" in general solution
  • Clustering Coefficient: Measure of the degree to which nodes in a graph tend to cluster together
    • Local version: Gives an indication of the embeddedness of single nodes
      • Local Clustering Coefficient = Number of links between vertices within its neighborhood/Number of links that could possibly exist between them
      • Clustering coefficient = Number of connections in the neighborhood of a node/Number of connections if the neighborhood was fully connected
    • Global version: Gives an overall indication of the clustering in the network
      • Clustering coefficient = Number of triangles connected to node i/Number of triples centered around node i = Number of closed triplets/Number of connected triplets of vertices = (3 × Number of triangles)/Number of connected triplets of vertices
      • Connected Triplet: A connected subgraph consisting of three vertices and two edges
  • Transitivity of Graph = 3 × Number of triangles in the network/Number of connected triples of nodes in the network
    • Reciprocity of Graph
      • In Directed network is the fraction of edges that belong to a loop of length two
  • EGO Network
    • Network consisting of a single node (ego) together with the nodes it is connected to (alters) and all the links among those alters
    • Diameter = 2
    • Size = Number of contacts an EGO has
    • Redundancy = Number of EGO alter/Size of EGO
    • Effective Size = Number of EGO alters – Sum of Redundancy of EGO alters
    • Efficiency = Effective size/Actual size
    • Weak components: Largest number of actors who are connected, disregarding the direction of the ties
    • Structural Hole: Gap between two individuals who have complementary sources to information
      • A hole between two contacts, provide network benefits to the third party
Share: