Inferential estimation of polypropylene melt index using stacked neural networks based on absolute error criteria

The modeling approach of stacked neural networks based on absolute error criteria is proposed, and applied to inferential estimation of polypropylene melt index. Single neural network model generalization capability can be significantly improved by using stacked neural network model. Proper determination of the stacking weights is essential for good stacked neural networks model performance, so determination of appropriate weights for combining individual neural networks based on absolute error criteria is proposed. For the purpose of comparison, single neural network models and stacked neural network models based on absolute error criteria are developed and evaluated. The application of the proposed modeling method to the development of melt index soft sensor in an industrial polypropylene polymerization plant demonstrates its effectiveness.