Probabilistic Rough Classification in Information Systems with Fuzzy Decision Attributes

Based on Pawlak’s two way approximations on Rough Sets and using thresholds G.Ganesan et al in 2004 proposed a method of rough indexing an information system which has fuzzy decision attributes. The limitation of Pawlak’s approximation is that it does not quantify the level of importance of the basic granules. Recently, Y.Y.Yao discussed Probabilistic Rough Set Model, which specified how basic granules could be quantified appropriately. In this paper, it is proposed to extend the work of G.Ganesan et al, taking into consideration the basic granule quantification mechanism of Probabilistic Set Model, thus generating more accurate rough indices for information systems with fuzzy decision attributes