Abstract: The variable precision rough set (VPRS) model based on dominance relations extends equivalence relations in traditional rough sets to dominance relations,and combines with the idea of variable precision to define the relevant concepts.Therefore,it can deal with preference-ordered information with certain fault tolerance degree.However,the definition of traditional dominance relation is still too strict,in which object x is superior to object y only when all attribute values of x are superior to that of y.This definition is difficult to be satisfied especially when the number of attributes is large.This will lead to smaller dominance,and even worse,it will affect the extraction of decision rules and then the process of decision making.To address this problem,the concept of dominance relation was extended by introducing a parameter and then the dominance set and approximation sets were correspondingly defined based on this extended do-minance relation.The extended VPRS model was also developed,and the coverage rate and the test accuracy were used as evaluation criteria to for model.Finally,an illustrative example was given and the experimens on UCI data were conducted to compare the proposed extended model with the traditional VPRS model.

Key words: Dominance relation,Variable precision rough set,Extended dominance relation,Approximation sets,Decision rules