Using mutual ranking probabilities for dimensionality reduction and ranking extraction in multidimensional systems of ordinal variables
Marco Fattore, Alberto Arcagni · BOA (University of Milano-Bicocca) · 2018
In this paper, we address the extraction of rankings from multi-indicator systems, as a problem of approximation between the so-called “mutual ranking probability” matrices, associated to the partial order relations derived from the data. After providing a theoretical treatment of the topic, we propose a practical algorithm for ranking extraction and show it in action on a real example, pertaining to regional competitiveness