In the field of multicriteria decision aid, considerable attention has been paid to supervised classification problems, especially to so-called sorting problems, where an order is assumed on the predefined classes. Recently, some non-supervised multicriteria classification procedures, also known as multicriteria clustering procedures, have been proposed aiming to discover data structures from a multicriteria perspective. We enlighten some properties of such approaches and their differences with regards to classical procedures, and we propose a taxonomy of this family of procedures. Moreover, we analyze extend to which these procedures differ from the multicriteria ranking problematic.
