THE RADIAL BASIS FUNCTIONS - PARTIAL LEAST-SQUARES APPROACH AS A FLEXIBLE NONLINEAR-REGRESSION TECHNIQUE

A new approach founded on Radial Basis Functions (RBF) and Partial Lea st Squares (PLS) is proposed to model non-linear chemical systems. Its performance is demonstrated for two simulated examples and compared w ith those of Multilayer Feedforward Network (MLP), Radial Basis Functi on Network (RBFN), and Spline-PLS. Good performance and a guaranteed l earning algorithm of the RBF-PLS approach makes it an attractive alter native for the earlier established methods.