[freebird]

Neural Network (NN) is a common way to estimation during aerodynamic optimization because of power of training in complex nonlinear problems, but it is need too input data at first that take too time.
I found same methods (listed below) but cant conclude that which of them is good for aerodynamics problems.

1- Response Surface Models (RSM)
2- Kriging method
3- Polynomial Regression(PR)
4- Multivariate Adaptive Regression Splines (MARS)
5- Gaussian Processes regression (MARS)
6- Cokriging
7- Radial Basis Functions(RBF)
8-Support vector machines(SVM)
9-Ensemble methods


Any suggestions/references about advantages or disadvantages of above methods would be greatly appreciated