The recognition of imbalanced classes is not an easy task for classifiers. Imbalanced classes are classes that are considerably smaller than other classes but not necessarily small ones. Most often smaller classes are more interesting from the user’s point of view but more difficult to be derived by a classifier. In this paper, which is a continuation of our previous works, we discuss a classifier using some inherent features of Atanassov’s intuitionistic fuzzy sets (A-IFSs, for short) making them a good tool for recognizing imbalanced classes.We illustrate our considerations on benchmark examples paying attention to the behavior of the classifier proposed (several measures in addition to the most popular accuracy are examined). We use a simple cross validation method (with 10 experiments). Results are compared with those obtained by a fuzzy classifier known as a good one from the literature.We also consider a problem of granulation (a symmetric or asymmetric granulation, and a number of the intervals used) and its influence on the results.
CITATION STYLE
Szmidt, E., Kacprzyk, J., & Kukier, M. (2016). Recognizing imbalanced classes by an intuitionistic fuzzy classifier. Studies in Fuzziness and Soft Computing, 332, 233–247. https://doi.org/10.1007/978-3-319-26302-1_15
Mendeley helps you to discover research relevant for your work.