CALIBRATION OF NONLINEAR ANALYTICAL SYSTEMS BY A NEURO-FUZZY APPROACH

Building in of fuzzy logic into the neural network structure and appli cation of the back-propagation learning algorithm allows construction of the neuro fuzzy system (NFS). The performance of this approach in c alibration of the non-linear systems was evaluated using simulated and experimental data sets. A comparison of the obtained results with an outcome of several other approaches, i.e., neural networks, and princi pal component analysis-neural networks, was performed.