Three term attribute description of Atanassov's Intuitionistic Fuzzy Sets as a basis of attribute selection

Abstract

We propose here a new proposal for attribute selection in the models expressed by the intuitionistic fuzzy sets. We further develop our previous paper in which the approach was already extended and the first computational tests were performed, i.e., the method was compared with the Principal Component Analysis (PCA). Here we test how the method behaves in comparison with the selection while using the Gain Ratio. We consider classification problems and try to reduce the number of attributes to not obtain substantially worse results.