We distinguish between directed and undirected, as well as between weighted and binary networks. The input networks are represented by their square connectivity (adjacency) matrices. Hence, the number of rows or columns corresponds to the number of nodes (vertices, units) in the network. A link (edge, connection) between any units i and j is denoted by a non-zero positive (i,j) entry -- correspondingly, the absence of a link is denoted by (i,j) = 0. For binary networks, all connections take the value (weight) of 1. For undirected networks, (i,j) = (j,i), for all i, j. For all networks, self-self connections are not allowed -- that is, (i,i) = 0, for all i.

The following table illustrates the hierarchy of network types. Note that any network type is a special case of all the types "above it" (e.g. a binary network is a special case of a weighted network). Consequently, a metric designed to work for any given network type, should also work for all the types "below it", but not, in general, for types that are on the same level or above.

Table: The hierarchy of network types.