This work proposes a novel approach that integrates multiple criteria ordinal classification with multiple criteria ordered clustering. Unlike traditional methods, which rely solely on predefined norms about classes or establish clusters based solely on relative comparisons, the proposed approach combines both perspectives. It employs absolute and relative comparisons based on indifference, preference, and incomparability relations, together with cluster consistency optimization to ensure that the resulting clusters are both internally homogeneous and externally well-differentiated according to the decision maker's (DM) preferences. The main innovative features of the proposed approach are: 1) indirect elicitation of preference model parameters: instead of requiring direct specification, parameters are dynamically adjusted within a feasible range defined by the DM, which enhances robustness and reduces sensitivity to specific parameter values; 2) compatibility with any preference model that validates the assertion ¿x is at least as good as y¿; 3) indifferent actions cannot be assigned to different clusters, thus ensuring cluster consistency; 4) the set of clusters satisfies monotonicity with respect to dominance and strict preference relations; 5) a metaheuristic optimization algorithm improves cluster consistency while maintaining clear separation between clusters. To demonstrate the effectiveness of the proposed approach, we applied it to two case studies: the classification of OECD governments and the Human Development Index (HDI). The results showed high consistency and interpretability, suggesting that this eclectic method can effectively overcome the limitations of most existing multiple criteria ordered clustering techniques.
