Premium
Process modeling using stacked neural networks
Author(s) -
Sridhar Dasaratha V.,
Seagrave Richard C.,
Bartlett Eric B.
Publication year - 1996
Publication title -
aiche journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.958
H-Index - 167
eISSN - 1547-5905
pISSN - 0001-1541
DOI - 10.1002/aic.690420913
Subject(s) - artificial neural network , process (computing) , generalization , computer science , nonlinear system , process modeling , stacking , artificial intelligence , machine learning , work in process , engineering , mathematics , chemistry , mathematical analysis , operations management , physics , organic chemistry , quantum mechanics , operating system
A new technique for neural‐network‐based modeling of chemical processes is proposed. Stacked neural networks allow multiple neural networks to be selected and used to model a given process. The idea is that improved predictions can be obtained using multiple networks, instead of simply selecting a single, hopefully optimal network, as is usually done. A methodology for stacking neural networks for plant‐process modeling has been developed. This method is inspired by the technique of stacked generalization proposed by Wolpert. The proposed method has been applied and evaluated for three example problems, including the dynamic modeling of a nonlinear chemical process. Results obtained demonstrate the promise of this approach for improved neural‐network‐based plant‐process modeling.