Ranking features by means of a qualitative optimization process

A set of qualitative measures describing a retail firm is considered. A qualitative ranking process of these features based on the managers' evaluations is presented. In a decision making context, this paper proposes a methodology to synthesise information given by ordinal data. Features are evaluated by each manager on an ordinal scale with different levels of precision and from two points of view: importance of the measure and performance on the measure. A representation for the different evaluations by means of k-dimensional qualitative orders of magnitude labels is proposed. The presented methodology deals with these evaluations to obtain two rankings based on comparing distances against a reference k-dimensional label. The final ranking of features with respect to their importance will help the managers to make decisions to improve the performance.