Intelligent Parameter Reduction Using Rough Sets Theory and Sensitivity Analysis

摘要

In this paper we propose an intelligent method based on the concept of rough sets theory to discover important parameters in an information system. The proposed method is presented as an algorithm which developed to detect non-important parameters and omitting them from the list of parameters. This method has been applied to the real problem of finding most affecting parameters influencing two hydraulic-geotechnical control variables in design of rubble mound breakwaters. On the other hand an intelligent sensitivity analysis employing the artificial neural network notions is performed on the parameters to compare with outcomes resulted from rough set based algorithm. Comparing results from these two approaches show that the rough sets approach has better performance.