Applied Research on Dynamic Modeling Method of Fermentation Process Based on SVM

For the difference and uncertainty between each batch of fermentation process, and currently the established models based on SVM are pre or off-line model, so once production conditions change, existing models may not be able to adapt to the new environment inevitably. And generalization capability of the model based on global learning support vector machine is not strong, so according to local learning theory the method of establishing the fermentation process dynamic model is proposed in this paper. The dynamic of the fermentation process model is realized through establishing the fermentation process dynamic sample sets. Taking the process of Escherichia coli producing interleukin-2 for example, experimental results verify that the method can establish a more accurate prediction model in the case of a smaller number of samples. Compared with the static SVM model, the dynamic model has a higher accuracy and a better dynamic adaptability.