About MC-SC2
Advisory Board
Useful Links
Assignment example



This website is about:

- Sorting problems;
- Classification problems; and
- Clustering problems.

Click for animation

The interest in MC_SC2 in the MCDA group was rather weak, comparing to Multiple Criteria Choice and Ranking. In contemporary decision practice it appears, however, that the Russell's saying is true: "in fact all decision is a classification", or even more, "knowledge is an ability of classification" [Roman Słowiński].

Groups can be defined a priori or a posteriori. In the latter case, it is common to build a procedure based on a similarity measure which assigns to the same group (cluster), objects having identical characteristics according to that measure. Groups are thus formed a posteriori. In literature these procedures are known as clustering techniques.

In the former case, groups (classes or categories) are pre-defined by using limit profiles, upper and lower bounds, central objects or other norms. Based on these reference objects a procedure is defined for assigning objects to the pre-defined groups.

When there is no preference relation among the groups (classes) we refer to the nominal classification (or discrimination) problem.

If the groups (categories) are ordered from the worst to the best one, or vice-versa, the problem is called an ordinal classification (or sorting or segmentation) problem.


[About the Authors]


[Home] [About MC-SC2] [Researchers] [Advisory Board] [Software] [MC-SC2 Bibliography][Forum] [FAQ] [Glossary] [Useful Links] [Contacts] [Acknowledgments]  



©2004-2008 by Juscelino ALMEIDA DIAS and Josť Rui FIGUEIRA.