采用RBF神经网络的状态监测数据趋势预测

State Monitoring Data Tendency Forecast Based on RBF Neural Network

采用RBF神经网络的状态监测数据趋势预测,通过选取状态参数、数据预处理、运用Matlab神经网络工具箱建立RBF神经网络模型。先对网络初始化,确定输入、输出和隐含层的节点数。再将网络输出结果与样本比较,根据求得误差值逐步调整隐含层神经元数量,直至误差满足实际需求为止。对网络仿真证明该法具有较高精度。

State monitoring data tendency forecast based on RBF neural network, through choosing state parameters, data pre-treatment and using the Matlab neural network toolbox, the RBF neural network model was established. At first, the network was initialised to ensure the node numbers of input layer, output layer and implication layer. Then, the network output results were compared with the sample; according to the error value, the nerve cells in implication layer was not adjusted gradually, until the error met the practical needs.The network simulation proved that this method was precisely.