基于支持向量机-遗传算法灰树花发酵模型的建立及优化

Establishment and Optimization of a Predictive Model for the Growth and Exopolysaccharide Production of Grifola frondosa Based on Support Vector Machine and Genetic Algorithm

对食用药用真菌灰树花发酵进行建模,获得使目标发酵产物达到最大产量的培养条件。运用支持向量机(support vector machine,SVM)方法进行非线性拟合,并采用遗传算法预测优化培养基成分,结果表明其能够较好预测灰树花发酵过程。运用此方法可在灰树花发酵生产过程中根据所需产物控制发酵条件与时间,具有较高指导意义。

To obtain the best medium constituents and culture conditions for maximum production of exopolysaccharides(EPS) by Grifola frondosa, nonlinear fitting was done using support vector machine(SVM) and the response variables, EPS production and mycelial biomass, were predicted using genetic algorithm. The results showed that the nonlinear model performed well in predicting the growth and EPS production of Grifola frondosa. The approach proposed in this study can provide a significant guideline to control culture conditions and time for the production of desired products by Grifola frondosa.