In this paper, the driving forces are represented by linear and nonlinear principal components. In order to extract such forces, dimension reduction is applied using spectral decomposition technique. The most contributing lower order principal components are retained to represent the driving forces. The modeling of these forces is achieved in comparison to the Gaussian one.
