FUZZY PNN ALGORITHM AND ITSAPPLICATION TO NONLINEARPROCESSES
A fuzzy polynomial neural network (PNN) is proposed. The fuzzy PNN is aimed at modelling high order nonlinear systems. The new technique fuses the conventional PNN by replacement of each neuron with fuzzy implication rules. In fact, each node of the PNN is operated as a small fuzzy system. The fuzzy membership of a node is expressed by Gaussian functions obtained by heuristics. The consequence of a node is expressed by constants and by regression polynomials. Optimal values of consequence parameters are selected by a least square method. The numerical simulation of a gas furnace and the \(\text{NO}_x\) emission of a gas turbine power plant shows that the proposed fuzzy PNN technique gives a more accurate prediction than conventional PNN and other methods of fuzzy modelling.
- A hybrid fuzzy model in nonlinear system modeling
- HYBRID FUZZY POLYNOMIAL NEURAL NETWORKS
- scientific article; zbMATH DE number 1192956
- Fuzzy polynomial regression with fuzzy neural networks
- Self-organizing polynomial neural networks based on polynomial and fuzzy polynomial neurons: Analysis and design.
- Fuzzy neural nets with non-symmetric membership functions and applications in signal processing and image analysis.
- A hybrid fuzzy model in nonlinear system modeling
- Power peak factor estimation using adaptive neural fuzzy inference system
- Fuzzy data processing using polynomial bidirectional hetero-associative network
This page was built for publication: FUZZY PNN ALGORITHM AND ITSAPPLICATION TO NONLINEARPROCESSES
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2747504)