Through the learning of experimental data, the neutral network model could be used to predict the shrinkage of large CPUE by taking the process parameters as input and product shinkage as output.

• 通过对实验数据的学习，利用该神经网络模型可以实现以工艺参数为输入，制品收缩率为输出的大型聚氨酯弹性体收缩率的预测。