A statistical approach to mining the DM strategy for OWA operators

The most important thing in using the ordered weighted averaging (OWA) operator which can easily describe the mental model of an Decision Maker (DM), is to characterize OWA weights. Determination of OWA weights cannot provide a characterization by itself. If we want to generalization and reusability of the OWA weights to aggregate various sized objects, we have to be determine more general form. In this paper, we propose a new approach for learning a stress function, which can be characterized as a DM strategy of OWA operator. For this aim the Kolmogorov-Smirnov test for similarity probability density functions is used.