Multi-Adjoint Fuzzy Rough Sets
Chris Cornelis, Jesús Medina, Nele Verbiest · Ghent University Academic Bibliography (Ghent University) · 2011
In this paper, we present an extension of the well-known implicator/t-norm based fuzzy rough set model based on a family of adjoint pairs. Rather than using a fixed implicator/t-norm pair to compute lower and upper approximations, our model allows to differentiate these fuzzy logical connectives according to the objects used in the calculation. We verify mathematical properties of the model, relate it to previous work in multi-adjoint property-oriented concept lattices, and discuss its possible use in data reduction.