| 切片平均分子量神经网络预测模型的研究及应用 |
| |
| 引用本文： | 卢云许,陆宝春,张世琪.切片平均分子量神经网络预测模型的研究及应用[J].南京理工大学学报(自然科学版),2000,24(2):139-142. |
| |
| 作者姓名： | 卢云许 陆宝春 张世琪 |
| |
| 作者单位： | 南京理工大学制造工程学院,南京,210094 |
| |
| 摘 要： | 建立聚合反应切片平均分子量的预测模型对锦纶帘子布的生产有重要的意义，该文采用改进的遗传算法（ＧＡ）和ＢＰ算法相结合的混合学习算法来训练神经网络，并采用多元逐步回归法对输入层节点数进行了优化，建立了聚合反应切片平均分子量在线预测的神经网络模型。在某化工厂聚合反应中的应用表明，该模型比基于最小二乘法的预测模型收敛速度快，预测精度高、网络有泛化能力强。
|
| 关 键 词： | 神经网络 多重回归 回归分析 遗传算法 |
| 修稿时间： | 1999－05－05 |
The Research and Application of Neural Network Predicting Model of Molecule Weight of Casting Slice Belt |
| |
| LuYunxu,LuBaochun,ZhangShiqi.The Research and Application of Neural Network Predicting Model of Molecule Weight of Casting Slice Belt[J].Journal of Nanjing University of Science and Technology(Nature Science),2000,24(2):139-142. |
| |
| Authors: | LuYunxu LuBaochun ZhangShiqi |
| |
| Abstract: | Polymerization is a complicated and important chemical process in the polyamide fibre production.It's significant to build the predicting model of molecule weight of casting slice belt to guide the production. The improved hybrid genetic algorithm and backpropagation algorithm are combined to train neural network, and the node numbers of input layer are optimized based on multiple stepwise regression. An on line predicting neural network model of molecule weight of casting slice belt in polymerization reaction is presented.Its application in certain chemical factory shows that this model has faster convergence,higher prediction accuracy and more network generalization than those of the least square. |
| |
| Keywords: | neural network multiple regression structural optimization genetic algorithm backpropagation algorithm on line predicting model |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
