Nonlinear system input structure identification based on rough set data analysis

It is important to identify the significant inputs of nonlinear system with large number of possible inputs before any known nonlinear modeling techniques can be applied. We propose an identification method based on rough sets data analysis. The method is driven only by raw data without assuming external structural information such as probability distributions or fuzzy membership functions. The information measure and performance of noise rejection are analyzed. Finally, two examples are given to validate the effectiveness of the method.