基于ISSA-XGBoost的毕赤酵母菌发酵软测量  

Soft measurement of pichia pastoris fermentation based on ISSA-XGBoost

作　　者：沈瑶 张立刚 王建扬 SHEN Yao;ZHANG Ligang;WANG Jianyang(College of Electrical and Information Engineering,Jiangsu University,Zhenjiang 212013,China;Yangzhong Victor Bioengineering Equipment Co Ltd,Yangzhong 212200,China)

机构地区：[1]江苏大学电气信息工程学院,江苏镇江212013 [2]扬中市威柯特生物工程设备有限公司,江苏扬中212200

出　　处：《传感器与微系统》2024年第8期122-125,共4页Transducer and Microsystem Technologies

摘　　要：针对毕赤酵母菌发酵过程菌体浓度难以在线检测,离线测量又存在极易染菌导致数据集不完整等问题,提出了一种基于改进麻雀搜索算法(ISSA)优化极致梯度提升(XGBoost)的软测量建模方法。首先,利用主成分分析(PCA)算法对样本数据进行主元分析,降低噪声和冗余度;然后,在标准麻雀算法(SSA)中引入自适应超参数和混合变异策略,增强了算法跳出局部极值和全局搜索的能力;最后,构建菌体浓度的ISSA-XGBoost软测量模型,并与XGBoost、SSA-XGBoost模型进行比较。仿真实验结果表明:ISSA-XGBoost模型的均方根误差(RMSE)、平均相对误差(MRE)均比XGBoost、SSA-XGBoost模型低,且ISSA-XGBoost的决定系数(R^(2))更接近于1,说明预测精度明显优于改进前,能够满足对毕赤酵母菌发酵过程菌体浓度的实时测量。Aiming at the problems that in fermentation process of Pichia pastoris,it is difficult to detect in concentration of bacteria online,and offline measurement has the problems of incomplete datasets caused by easy infection of bacteria,a soft measurement modeling method based on improved sparrow search algorithm(ISSA)optimized extreme gradient boosting(XGBoost)is proposed.Firstly,the principal component analysis(PCA)algorithm is used to analyze on sample data to reduce the noise and redundancy.Secondly,adaptive hyperparameters and hybrid mutation strategy are introduced to standard sparrow algorithm(SSA)for enhancing the ability of the algorithm to jump out of local extremum and global search.Finally,the ISSA-XGBoost soft measurement model of cell concentration is constructed and compared with XGBoost and SSA-XGBoost models.The simulation experimental result shows that the root mean square error(RMSE)and mean relative error(MRE)of the ISSA-XGBoost model are lower than those of the XGBoost and SSA-XGBoost models,and the determination coefficient R^(2)of the ISSA-XGBoost model is closer to 1,indicating that the prediction precision is significantly better than that before the improvement.The demand for real-time measurement of bacteria concentration in Pichia pastoris fermentation process can be satisfied.

关 键 词：毕赤酵母 麻雀算法 极致梯度提升 软测量模型 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]