Create a function that calculates conductance based on the Teng et. al. paper:
To complement these standard network metrics, we employ
two additional measures. The first, cycle-free effective
conductance, defined in [12], captures the potential for directed
information flow between a pair of nodes i and j in
the network:
Cij =SUM(p)( PRODUCT(k,l)(w(k, l)/deg(k) ) )
where p is a path between nodes i and j, k and l are nodes
in the path with edge (k, l) of weight w(k, l) between them,
and deg(k) is the out-degree of node k. From this definition,
there will be high conductance from node i to node j if there
are many paths, if the paths are short, and if they go through
nodes that have few neighbors besides the ones on the path.
To obtain the conductance of the entire graph, we sum over
all pairs of nodes present in the time segment:
Create a function that calculates conductance based on the Teng et. al. paper:
Cij =SUM(p)( PRODUCT(k,l)(w(k, l)/deg(k) ) )