论文部分内容阅读
函数连接型神经网络是一种无隐含层的新型网络，应用其三阶联合激励增强特性对某矿区矿石与围岩进行判别研究，识别准确率近１００％．在对预测集的每一个输入信号添加２０％的噪音干扰后，发现依然能准确判别．可见网络的容错能力是十分令人满意的．
Functional connectivity neural network is a new type of network with no hidden layer. By using the third-order combined excitation enhancement characteristic, the discriminant study of ore and surrounding rock in a mining area is carried out, and the recognition accuracy is nearly 100%. After adding 20% of the noise to each of the input signals in the prediction set, it was found that accurate discrimination was still possible. Visible network fault tolerance is very satisfactory.