模糊聚类网中多对一映射的实现

Realization of Multitude to One Mapping for Fuzzy Clustering Neural Networks

分析了模糊聚类网实现多对一映射的两种方式：有监督和无监督，给出了实现这两种方式的网络模型：自适应模糊集成聚类网（ＩＡＦＣＭＡＰ）和二次模糊聚类网（ＤＵＡＬＦＣ）前者在有监督的学习中实现多对一的思想。

This paper analyses two methods of multitude to one mapping for fuzzy clustering neural networks:supervised and unsupervised.We construct two new network models:the integrated adaptive fuzzy clustering and mapping networks(IACMAP),the dual fuzzy clustering neural networks(DUALFC).The former realizes multitude to one mapping idea during supervised learning.The latter completes multitude to one mapping idea through dual clustering.Some experimental results demonstrate that the two networks are clearly superior to the popular fuzzy clustering networks.