Correlation coefficients in T-spherical fuzzy environment using statistical viewpoint and their applications

A T-spherical fuzzy set (T-SFS) is a generalisation of spherical fuzzy sets (SFSs), picture fuzzy sets (SFSs), intuitionistic fuzzy sets (IFSs), and fuzzy sets (FSs) in which the sum of the qth power of membership, the qth power of non-membership and qth power of neutrality values is at most one. The correlation coefficient is a crucial tool in fuzzy/non-standard fuzzy theory and has been applied in various fields such as clustering, pattern recognition, medical diagnosis, decision-making, etc. However, the existing correlation coefficients for T-SFSs give only the degree of correlation between two T-SFSs but do not tell us the nature of correlation (negative or positive). In this study, we propose two correlation coefficients for T-SFSs, which not only give the strength of correlation between two T-SFSs but also tell us whether the two T-SFSs are positively correlated or negatively correlated. We also discuss several properties of these correlation coefficients. We apply these correlation coefficients to solve a pattern recognition problem in the T-spherical fuzzy environment and compare the results with some existing measures. Further, by considering linguistic hedges, we theoretically and empirically contrast the performance of the proposed coefficients of correlation for T-SFSs with several existing measures.