Metric | Description |
---|---|
Degree centrality | The number of links directly connected to that player |
Betweenness centrality | v i is defined as follows. Consider the j th pair of players (v ′ , v ′′) j for which v′ ̸= v i and v′′ ̸= v i . Let T j be the number of shortest paths between v ′ and v ′′ . Let n j be the number of these paths that contain player i. The betweenness centrality of v i is \( {\displaystyle \sum_{j=1}^m\frac{n_j}{T_j}} \) where m is the number of player-pairs that connect by at least one path [35]. The measure incorporates by re-defining the length of a path between players v i and v j to be the sum of the link weight inverses across all of the links in the path [44]. A player with high betweenness centrality has control of information propagation in a network |
Eigenvector centrality | Let W be the weighted-link adjacency matrix. Let e be the first eigenvector of W. The eigenvector centrality of player v i is the i th component of e [35]. This metric measures a player’s influence by measuring how easily information can flow between a player and all other players regardless of the path taken [43] |
Network connectedness | The largest eigenvalue of W. Let CI be this index |
Gould-Fernandez total brokerage score | The GF total brokerage score for vi is the number of player pairs, v’, v” for which (v’, vi) and (vi, v”) are both in E but the link (v’, v’) is not in E [35]. A player with a high brokerage score functions as an intermediary (or broker) for many pairs of players not directly connected to each other |