摘要： 提出了一个基于模糊集理论的新的神经网络结构及其学习算法。这种神经网络是对标准的BP网络修改后构成的（记作FIBP）。并通过几个实例仿真验证其有效性。实验结果表明，当用FIBP网络解决动态的、高度非线性的函数逼近时，其学习速度比BP网络快，精度高而且泛化能力高。

关键词: 神经网络, 模糊集, 学习算法, 泛化能力

Abstract: This paper presents a novel neural network architecture based on fuzzy set theory, FIBP. Its architecture is constructed by modifying the standard BP network. Its learning algorithm is also proposed. The network is tested using several dynamic and highly nonlinear examples. Experimental results demonstrate that the fuzzy neurons in the network and its learning algorithm provide enhanced network architecture and improve the performance of these networks significantly.

中图分类号: 

• TP183