In traditional rough set theory, it always used indiscernible relation classify objects, but, if the decision table with continuous attributes, classify objects with indiscernible relation is too strictly that each object maybe form a class. In this paper, first, the authors generalize the indiscernible relation to similarity relation, give a definition of λ-discernibility matrix; then, according to the properties of λ-discernibility matrix, an attribute reduction algorithm of decision table with continuous condition attributes is put forward...
