论文部分内容阅读
讨论了回归神经元网络（ＲＮＮ）的网络结构和基本实现方法，提出了主元分析（ＰＣＡ）和具有自校正功能的回归神经元网络相结合的非线性时变系统预报建模方法，并用于减压塔塔顶温度的预报．结果表明，该方法具有良好的预报性能．
The network structure and basic realization method of RNN are discussed. A nonlinear time-varying system modeling method based on principal component analysis (PCA) and self-correcting regression neural network is proposed and applied to Prediction of tower top temperature of vacuum tower. The results show that the method has good forecasting performance.