A Flexible Multi-Attribute Classification Method for Incorporating Decision-Maker Preferences in the Decision-Making Process

Multi-attribute decision-making (MADM) plays a vital role in addressing classification problems across various fields. Despite numerous methods developed for evaluating alternatives, there is a research gap in the MADM approach for effectively incorporating decision-maker supplementary preferences into the classification process. Classical multi-attribute classification methods cannot consider decision-maker supplementary preferences. Not considering these data may make it difficult to use the classification results directly. This research proposes a Flexible Central Profile-based Classification Method (FCPCM) that not only allocates alternatives to classes but also incorporates decision-maker supplementary preferences into the classification process. This research has also proposed a simplified algorithm for problems that lack decision-maker preferences. The developed method is used in two case studies, including portfolio selection and classifying countries according to their level of economic freedom. In both case studies, the results of the FCPCM method were compared with those obtained from TOPSISSort-C and FlowSort methods. The results showed a high quality of classification with the developed method. The proposed approach in this research can be used for other real-world classification problems.