Stacked Neural Network Modeling Applied to the Synthesis of Polyacrylamide‐Based Multicomponent Hydrogels
Florin Leon, Ciprian George Piuleac, Silvia Curteanu · Macromolecular Reaction Engineering · 2010
Abstract Artificial neural networks are a convenient method of modeling processes for which the analytical estimation is difficult or computationally intensive. Besides the performance on training data, the generalization capability of the resulting neural model is very important, especially for problems such as the one described in this paper, the synthesis of polyacrylamide‐based multicomponent hydrogels, because the prediction of the swelling degree are extremely useful in practice, because it can replace the experiments that necessitate a great amount of materials and time. In order to increase the generalization performance, we use stacked neural networks with a procedure of weight optimization, which prove to be a good choice regarding the accuracy of the results provided by the model. magnified image