Titlebook: Artificial Intelligence and Soft Computing – ICAISC 2006; 8th International Co Leszek Rutkowski,Ryszard Tadeusiewicz,Jacek M. Żur Conferenc

手势

偏离

Harpoon

圆木可阻碍

RBF Nets in Faults Localizationwith a metric in the label space. Then, this problem is approximately solved, providing hints on selecting appropriate RBF nets. It was shown that the approximate solution is the exact one in several important cases. Finally, we propose the algorithm for learning the proposed RBF net. The results of

抚慰

A Hypertube as a Possible Interpolation Region of a Neural Models based on the parametric curve modelling. The idea of it is to surround the parametric curve model with the hypertube covering most of the data points used in a neural model training. The practical application of the method will be shown via a system of an unemployment rate in Poland in years 1992-

步履蹒跚

遵循的规范

Fast Orthogonal Neural Networksof basic operations associated with the algorithm of a given transform is used in order to substantially reduce the number of adapted weights of the network. Two new types of neurons corresponding to orthogonal basic operations are introduced and formulas for architecture-independent error backpropa

骨

alabaster

A Fast and Numerically Robust Neural Network Training Algorithm adaptively-adjustable time-varying forgetting factor technique, is presented first. Then a U-D factorization-based RPE (UD-RPE) algorithm is proposed to further improve the training rate and accuracy of the FNNs. In comparison with the backpropagation (BP) and existing RPE based training algorithms

Nonthreatening