Abstract: Melt index prediction of polypropylene based on stacked neural networks-ridge regression (SNNs-RR) was studied. Single neural network model generalization capability could be significantly improved by using the stacked neural network model. Proper determination of the stacking weights was essential for good SNNs model performance, the determination of appropriate weights for combining individual networks using ridge regression was proposed. The results of using SNNs-RR model demonstrated significant improvement in model accuracy and robustness, as compared with using the single neural network model.

Key words: 聚丙烯, 熔融指数, SNNs-RR, 软测量

摘要： 提出了一种组合神经网络-岭回归（SNNs-RR）建模方法，并将该方法应用于聚丙烯熔融指数软测量研究中.通过多个单一神经网络的合理组合可显著改善神经网络模型的泛化能力，而选择合适的组合权重对组合神经网络模型是否具有良好预测性能是至关重要的，因此提出了采用岭回归方法来选择合适的组合权重.通过与单一神经网络模型的预测结果进行比较，表明基于SNNs-RR的聚丙烯熔融指数软测量模型具有更佳的预测精度和鲁棒性.

关键词: 聚丙烯, 熔融指数, SNNs-RR, 软测量