Labeled Fuzzy Rough Sets in Multiple-Criteria Decision-Making

This paper presents a hybrid approach to constructing a decision-making process, basing on the concept of labeled fuzzy rough sets. We introduce a modified definition of fuzzy rough approximations which are necessary, when an increased value of the similarity threshold is chosen in the determination of fuzzy linguistic labels. Labeled fuzzy rough sets are used in the first stage of decision-making process for analyzing the decision system of an expert, who recommends objects by respecting the guidelines of a decision-maker. In the second stage, the linguistic labels obtained from the expert are applied for determining a final ranking of objects, according to preferences which are imposed by the decision-maker on fuzzy attributes and their linguistic values. A short example of analysis helps to elucidate the presented method.