基于三次样条权函数神经网络的钢坯温度预报

针对传统神经网络对铜坯温度预报存在的一些缺点，如：隐层数不易确定．网络训练对初值敏感等。该文利用三次样条权函数神经网络建立了钢坯温度预报模型．克服了传统神经网络的缺点。仿真结果表明该模型具有较高的精度。

Slab Temperature Prediction Based on Neural Networks with Cubic Spline Weight Function

Aiming at some disadvantages of the traditional neural networks on billet temperature prediction,such as: hard to fix the hidden layer units,sensitive to initial for networks training,etc,a kind of new billet temperature prediction model is established based on the cubic spline function neural networks.The disadvantages of the tradi-tional neural network are overcome,and the simulation results show that the model has higher precision.