Contents
- Index

**NETWORKS>CENTRALITY>MULTIPLE MEASURES**

**PURPOSE **Computes user selected centrality measures for binary data.

**DESCRIPTION **Produces a table of user selected centrality measures for both directed and undirected data. The measures are degree, eigenvector, Bonacich Power, K-step reach, Average Reciprocal Distance and betweenness. All of these are avilable individually in other routines but the following should be noted. The eigenvector routine for directed data has the option of calculating either on the symmetrized graph or on the directed graph. If the directed option is selected then both the right eigenvector (OutEigen) and the left eigenvector (InEigen) are calculated. (The standard eigenvector centrality routine uses symmetrized data but the eigenvectors for directed data can be calculated in the tools menu). The user can select whether normalized or raw scores are collected. The raw eigenvector scores are calculated such that the most central actor has a score of one, the normalized ones have a Euclidean norm equal to the size of the network.

**PARAMETERS** **Input Network:**

** **Name of file containing network to be analyzed. Data type: Digraph

**Output Measures:** (Default = <filename>-cent')

Name of file which will contain centrality measures for each node.

**Data are**

** **Select either directed or undirected or allow the routine to detect the nature. If auto is used then directed data with only recirocated ties will be treated as undirected.

**Report**

** **Select either raw or normalized. Note that non-normalized ARD is the sum of reciprocal distances.

**Measures**

** **User selects which measures to include. Bonacich power requires a beta value, the default will select the largest permissible value and this will be reported in the log file. K-step reach requires K to be specified and this is set as 2 by default. If directed data has been selected then the eigenvector method will be unchecked. Checking once gives a grey tick which means the eigenvector will only be found for symmetric matrices. If it is checked again it will turn black indicating that both left and right eigenvectors will be computed for non-symmetric matrices. If the auto-detect is selected then it will default to symmetric matrices only.

**LOG FILE **If the Bonacich power method was selected the value of beta will be given. If eigenvector was selected the principal eigenvalue will be reported followed by the table of centrality measures.

**TIMING** O(N^3).

**COMMENTS**

**REFERENCES** See individual measures.