We consider an arbitrary crisp domain where a resemblance relation is
defined, and we are going to extend it to work with fuzzy values. This
will lead us to the concepts of weak and strong resemblance. The exte
nsion may be carried out with or without the help of precision levels
for the fuzzy values. We show certain equivalence between both extensi
ons, and apply this study to the definition of fuzzy dependencies. (C)
1997 Elsevier Science B.V.