论文部分内容阅读
静态神经网络由于自身的局限性难于对非线性时变过程进行建模和预测, 而最小资源分配网络(M-RAN)又因调节参数过多难于实现. 提出了一种新型的基于局部投影概念的RBF网络序贯学习算法: 局部投影网络LPN, 进而对算法进行了最小化改进. 在此基础上进行了详细的算例验证.
Static neural networks are difficult to model and predict nonlinear time-varying processes due to their own limitations, but M-RAN is difficult to implement due to too many adjustment parameters. A new concept based on local projection RBF network sequential learning algorithm: local projection network LPN, and then the algorithm was minimized to improve.On this basis, a detailed example validation.