Attribute Reduction for Fuzzy Set Valued Information Systems

摘要

Rough set is considerably used to perform attribute reduction in an information system. In real life situations, several objects contain many values for an attribute. Information systems with such conditions are termed as set valued information systems. Initially, tolerance relation was used to handle set valued data. This tolerance relation was based on intersection of set values which can lead to loss of some information. Hence, fuzzy rough model is constructed for such information system which uses fuzzy relation. However, in real situations, there is a possibility that some of the set valued attributes are real or continuous. Hence existing set valued fuzzy rough model requires some changes to be done. In this paper, fuzzy relation for fuzzy set valued data is defined and has been used as a basis to find the superfluous attribute, rcduct and relative rcduct for fuzzy set valued information system.