Choosing suitable similarity measures to compare intuitionistic fuzzy sets that represent experience-based evaluation sets

Which similarity measures can be used to compare two Atanassov's intuitionistic fuzzy sets (IFSs) that respectively represent two experience-based evaluation sets? To find an answer to this question, several similarity measures were tested in comparisons between pairs of IFSs that result from simulations of different experience-based evaluation processes. In such a simulation, a support vector learning algorithm was used to learn how a human editor categorizes newswire stories under a specific scenario and, then, the resulting knowledge was used to evaluate the level to which other newswire stories fit into each of the learned categories. This paper presents our findings about how each of the chosen similarity measures reflected the perceived similarity among the simulated experience-based evaluation sets.