An Intelligent Data Mining Approach Using Neuro-Rough Hybridization to Discover Hidden Knowledge from Information Systems

In this paper we discuss on the necessity of applying data mining operators on information systems containing a set of variables which describe the characteristics and behaviors of a specific system and could be exploited in approximating system’s functionality. For the problem of function approximation, we developed a new approach combining two intelligent methods. At first we used an algorithm based on the notions of rough set theory as a preprocessor to our information system. Afterward an artificial neural network is employed as a function approximator to obtain values for decision attributes of information system while values of condition ones are passed to the network. This method has been applied to a real problem of approximating values for two hydraulic-geotechnical control variables of rubble mound breakwaters, and the results have been discussed.