基于KTA-LSSVM的青霉素发酵过程预测建模

为解决青霉素发酵过程预测建模中存在的输入变量选择问题，提出了基于核目标度量（kernel target alignment，KTA）和最小二乘支持向量机（least squares support vector machines，LSSVM）的青霉素发酵过程预测模型。首先，在分析影响青霉素产物浓度相关因素的基础上选取输入变量，采用KTA对输入变量进行尺度缩放，然后，利用Pensim仿真平台数据，采用混沌粒子群算法对LSSVM的参数寻优，建立青霉素发酵过程的KTA-LSSVM预测模型。青霉素浓度预测的KTA-LSSVM模型均方根误差为0.0179，LSSVM模型的均方根误差为0.0276，实验结果表明，本文提出的模型预测精度高，推广性能好。

Prediction model of penicillin fed-batch fermentation based on KTA-LSSVM

A new prediction method of penicillin fed-batch fermentation based on kernel target alignment （KTA） and least squares support vector machines （LSSVM） was proposed to deal with input variable selection. Firstly，input variables were selected by analyzing the factors which affected penicillin concentration，then，the KTA feature rescaling method was used to rescale input variables. Finally，the simulation data from Pensim simulation platform was used to establish the LSSVM model in the penicillin fed-batch fermentation by using chaos particle swarm optimization （CPSO） on the LSSVM parameters optimization. The root mean square error （RMSE） of the proposed method was 0.0179，whereas the RMSE of LSSVM was 0.0276，showing better prediction and generalization.