论文部分内容阅读
为克服BP神经网络模型及其学习算法中的固有缺陷,构造了第二类Chebyshev前向神经网络模型,提出该神经网络模型权值直接确定法和结构自适应确定法.理论分析及仿真实验均表明,该系统弥补了BP神经网络的某些固有缺陷.相比同构型BP神经网络,其计算速度和工作精度均有大幅提高.
In order to overcome the inherent flaws in BP neural network model and its learning algorithm, the second Chebyshev neural network model is constructed, and the direct determination method of weights and the adaptive structure determination of the neural network model are proposed. The theoretical analysis and simulation experiments It shows that the system makes up for some inherent flaws of BP neural network, and its computational speed and work precision are greatly improved compared to BP neural network with isomorphism.