Rough Set Approximations: A Concept for Incomplete Information System

Rough set theory is based on the concern of equivalence classes or a given training data. Rough set theory can be used in classification to find relationships and noisy in data. Data tuples forming an equivalence class are indiscernible; the samples are identical with respect to the attributes describing the data. Rough sets used the concepts of approximately or “roughly� define classes. The lower approximation consists of all the data that, is based on the knowledge of the attributes and certain to belong without ambiguity. The upper approximation consists of all the data that, is based on the knowledge of the attributes and cannot be described. The RST are applied in several fields including image processing, data mining, pattern recognition, medical informatics, knowledge discovery and expert systems. Rough set have been combined with methods such as neural networks, fuzzy logic etc resulting some good results. In this paper we used the concepts and properties of rough set theory to handle incomplete information. We construct missing values with the help of rough set properties. We found that rough set theory recovered missing values more efficiently as compared to other missing techniques.