样条权函数神经网络是一种新型的神经网络，它克服了传统神经网络收敛速度慢、初值敏感、局部极小的问题。因其能精确学习给定的样本，并且结构简单、训练速度快，因此被广泛关注。结合分子三次、分母一次的有理样条函数和样条权函数神经网络的优势，研究了分子三次、分母一次乘性有理样条权函数神经网络，并对其灵敏度进行了理论分析和实验仿真。通过理论分析和仿真可以看出，该神经网络具有分子三次、分母一次的有理样条和样条权函数神经网络的优越特性，在一定扰动范围内，该样条权函数神经网络的灵敏度稳定，具有很强的抗干扰能力。%The neural network with spline weight function is a new kind of neural network,which overcomes many problems such as slow convergence speed,sensitive to initial value and local minima. It is widely concerned because of its accurate learning approach to given patterns,simple network topology,fast training speed and so on. Based on the advantage of neural network with spline weight function, the sensitivity of neural network with cubic numerator and linear denominator of rational spline weight functions using multiplicative neu-rons is discussed,and the accuracy of analytical results is verified by simulation. Both the theoretical analysis and simulation results show that when the disturbance is in a certain range,the sensitivity of this kind of spline weight function neural network is very stable,and is featured with strong noise resistance.
展开▼
