Extension of rough sets model of generalized information system

In order to improve the ability of dealing with imprecise linguistic terms in information system, the paper study the extended rough set model of generalized information systems. The paper defines a kind of generalized information system and extends almost all basic concepts of rough set theory to fit this case. The paper has also considered the construction of homogeneous granules by introducing the T_L-similarity property and the principle of Ziarkopsilas variable precision to the fuzzy case. The reduction approach proposed in the paper can ensure all important information and only efficiently eliminate that not essential ones, which is superior to the Pawlakpsilas typical approach in the simplification of decision tables.