Relevance in preference structures

Fuzzy preference and aversion relations allow measuring in a gradual manner the attitude of the individual regarding some pair of alternatives. Following the Preference-Aversion (P-A) model, previously introduced for identifying the subjective cognitive state for some decision situation; here, we explore a methodology for learning relevance degrees over the complete system of alternatives. In this way it is possible to identify in a quick way, the pieces of information that are more important for solving a given decision problem.