Optimal scoring of partially ordered data,with an application to the ranking of smart cities
In this paper, we propose a solution to the problem of scoring and ranking partially ordered data, by exploiting the spectral properties of so-called matrices of mutual ranking probabilities, a class of matrices which comprise and convey information on the dominance among statistical units. The procedure is optimal in many respects and overcomes the limitations of other ranking tools, which may fail to deliver acceptable solutions, even in trivial cases. We show the algorithm in action, on real data pertaining to the smartness of a subset of European cities.