期刊名称：International Journal of Hybrid Information Technology
印刷版ISSN：1738-9968
出版年度：2016
卷号：9
期号：9
页码：13-22
出版社：SERSC
摘要：In this paper a dynamic variation of neural network system was presented and used to chemical reaction identification. Proposed system is consists of two parts. First one is using first principles rules, which infers a priori knowledge. The other one is a variation of backpropagation neural network algorithm, which estimates parameters of chemical reaction system. This kind of system is useful for chemical modeling processes. Merit of proposed system is once it have learned, it can be used for process control and optimization. Proposed chemical reaction inference system is a powerful and estimates well on process parameters.
关键词：Backpropagation algorithm; Parameter estimation; Hybrid system
