i.e. the elements of each column sum up to 1, so the matrix is a stochastic matrix (connaissance more details see the computation case below). Thus this is a variant of the eigenvector centrality measure used commonly in network analysis. Parmi fait c'levant très simple, Celui-là