头孢菌素C；_反向传播；_神经网络；_预测_

利用神经网络预测头孢菌素C的生物合成

http://www.100md.com 《中国抗生素杂志》 2007年第1期

头孢菌素C；,反向传播；,神经网络；,预测,,头孢菌素C；,反向传播；,神经网络；,预测,1材料与方法,2结果与分析,3结论,参考文献

     摘要：利用头孢菌素C发酵过程积累的数据，建立BP神经网络预估模型，实现以发酵前期的菌浓和pH对效价的预测，将此模型应用于生产实际，分别通过倒种和改变培养基中碳源组成的方法，使头孢菌素C的合成水平分别提高了118%和157%，表明模型具有较好的预测功能。

    关键词：头孢菌素C；反向传播；神经网络；预测

    Prediction of the cephalosporin C biosynthesis by a back propagation neural network model

    Ji Zhixia, Chu Ju, Zhuang Yingping and Zhang Siliang

    (State Key Laboratory of Bioreactor Engineering ECUST, Shanghai 200237)

    ABSTRACT The back propagation (BP) neural network model was set up by cephalosporin C fermentation data, and the productivity was forecasted by PMV and pH changing tendency of early fermentation phase. The model was proved to be effective and with good prediction capacity. By increasing inoculum and changing medium carbon source, the productivity of cephalosporin C fermentation was further increased by 118% and 157%, respectively, which was well predicted by the established model.

    KEY WORDS Cephalosporin C; Back propagation; Neural network model; Prediction

    实现发酵生产过程的优化操作和控制，单凭经验或经典的试验数据是无法满足要求的，因此有必要建立模型指导发酵过程的优化，一些难于在线检测的重要变量、如菌体浓度，底物浓度和产物浓度等可借助于已建立的数学模型，通过测量与其相关的其它可在线测量的变量 ......

百拇医药网 http://www.100md.com/html/DirDu/2007/04/28/42/80/84.htm

您现在查看是摘要页，全文长 10959 字符。