1303 Feature Extraction from Design Space Using Rough Sets
Details
Data mining technique based on Rough Set theory has been applied to solutions obtained from three-objective optimization for regional jet wing design with engine nacelle. To reveal design tradeoffs, multiobjective optimization was performed by using Evolutionary Algorithms coupled with the Kriging model. High dimensional data (design parameters and the corresponding objective function values) are mapped onto the two-dimensional Self-Organizing Map (SOM) where global tradeoffs are visualized. The rule sets are derived by Rough Set theory so as to determine the importance of design parameters corresponding to the clusters obtained from SOM.
