Given asymmetric distances are a hard problem to treat explicitly in any context, including clustering. The usual way out is to somehow symmetrise hem at some point of the procedure. Even though, for quite fundamental reasons, asymmetric structures can hardly be analysed “by inspection” for even modest dimensionality of problems, it is argued in this paper that fuzzy clustering is one domain, in which asymmetric distances can be accommodated in a quite natural manner. Following the previous paper of the author, a procedure is proposed, here complemented with additional analysis, emphasis being placed on the quality criterion (objective function).
Operational Program Digital Poland, 2014-2020, Measure 2.3: Digital accessibility and usefulness of public sector information; funds from the European Regional Development Fund and national co-financing from the state budget.
Oct 15, 2021
Jul 19, 2021