摘要
采用结构简单的 RBF 神经网络建立真菌发酵过程模型,模拟结果精确,网络训练误差收敛速率也很快。
The suspended cultrue of fungus is quite a complex biochemical process.It is difficult to simulate this process very precisely with normal dynamic model.Apllying extended kalman filtering algorithm was able to improve the simulation to a certain extent,but that did not arrive at good condition.In this paper,the simplest artificial neural networks—RBF Networks was used to model the process of fungus fermentation and very precise simulations were achievable.Also the training error converging rate was high.
出处
《苏州医学院学报》
1996年第3期411-413,598,共3页
Acta Academiae Medicinae Suzhou
关键词
真菌发酵
神经网络
Fungus fermentation
RBF Networks
modeling
