Title :
Toward accurate Choquet integral-based neural network
Author :
Yue, Shi-hong ; Wang, Zheng-you
Author_Institution :
Sch. of Electr. Eng. & Atomization, Tianjin Univ., China
Abstract :
A new Choquet integral-based neural network is presented. In light of examining the structure characteristics of conventional neural networks, the initial size of the new neural network is determined, and then using the pruning algorithm of general neural network the final size can be done. Therefore, the new neural network can work more efficient compared with the conventional Choquet integral-based neural networks. Two experiments are used to verify the validation of the new neural network.
Keywords :
fuzzy neural nets; fuzzy set theory; integral equations; CFI NN; Choquet fuzzy integral; Choquet integral-based neural network; fuzzy set theory; pruning algorithm; Electronic mail; Feeds; Finance; Fuzzy neural networks; Fuzzy sets; Handwriting recognition; Information technology; Neural networks; Neurons; Open wireless architecture; Fuzzy integral; neural network; pruning algorithm;
Conference_Titel :
Machine Learning and Cybernetics, 2005. Proceedings of 2005 International Conference on
Conference_Location :
Guangzhou, China
Print_ISBN :
0-7803-9091-1
DOI :
10.1109/ICMLC.2005.1527753