Optimization Of Phellinus Igniarius Culture Environment Based On Artificial Neural Network

Flavonoid is antioxidation and anticancer,and it also has enormous medical value,whose demand is great especially in the research of its medical use.However,it is rare in natural environment,thus the number of Flavonoid extraction from natural Phellinus Igniarius can not meet the needs of medical research.Flavonoids are secondary metabolites from liquid fermentation and biological experiments showed that flavonoid yield is mainly related to the culture conditions of Phellinus Igniarius,leading optimized culture environment problem to be an important issue in flavonoids extraction.It is believed that biological experiments are unrepeatable,complex and diverse.Exceptional data point and missing data are always along with uncontrollable factors during experiment process.A detection clustering algorithm for exceptional data point based on distance and an interpolation method based on statistics and the equation of regression are proposed for data expansion and pretreating Phellinus Igniarius liquid fermentation experiment data,providing the premise conditions.Backpropagation Neural Network is a nonlinear optimization model by continuous training,which is used to simulate multidimensional nonlinear models,and has been applied in many engineering fields.According to the liquid fermentation of Phellinus Igniarius data,principal component analysis algorithm based on network topology optimization method is proposed,which realizes network input node merging and solves BP neural network model training unstable and long training time problem,and also improves the accuracy of prediction model.In this paper,a new optimization algorithm combining particle swarm algorithm and genetic algorithm is proposed based on a new algorith,and an optimization model of the environment of liquid fermentation of Phellinus Igniarius culture is estiblished based on improved BP neural network model.Experimental analysis shows that network topology optimization method based PCA can effectively improve the training sp,eed and prediction accuracy of BP neural network model.Besides,the stability and efficiency of the model we better.Compared with two basic algorithm,the new one effectively reduces the space complexity and time complexity.Through data from simulation experiment,the optimum culture conditions of liquid fermentation of Phellinus Igniarius are found and real flavonoids production has been significantly improved,providing constructive suggestions for the next step of the biological experiment.At the same time,it opens a new door for solving other biology problems.