An open-source software package to assess similarity measures that compare intuitionistic fuzzy sets

Since intuitionistic fuzzy sets (IFSs) have been applied to solve problems in topics like decision-making or pattern recognition, the study of similarity measures aiming to compare this kind of fuzzy sets has become a challenging research subject. When proposed, a similarity measure is usually tested to demonstrate its properties and advantages over the others. However, those tests are occasionally performed using a small number of examples that do not allow a researcher or practitioner to detect potential drawbacks. In this paper, an open-source software package whereby a researcher can empirically assess several (configurations of) similarity measures while comparing IFSs that characterize experience-based evaluations is proposed and presented. By means of the proposed package, one can (1) build a large number of IFSs according to different learning scenarios, (2) compare those IFSs using existing or novel similarity measures, and (3) generate a comprehensive report about how each similarity measure reflects a perceived similarity. Reports generated by the package show that only a few of the existing similarity measures reflect properly a perceived similarity when IFSs resulting from opposite learning scenarios are compared to each other.

ISBN:9781509060344

Publication year:2017

BOF-keylabel:yes

IOF-keylabel:yes

Accessibility:Closed