基于SVM的生物离心机工艺参数优化研究

Study on Optimization of Process Parameters Based on SVM

为提高透析器离心机工艺参数精度,提出基于SVM的生物离心机工艺参数优化方法。通过高斯核函数RBF改进PSO-SVM算法,主要从核函数中心计算、核函数宽度计算和PSO算法优化SVM函数三个方面进行改进,由此取得RBF-PSO-SVM方法,通过该改进算法得到最优工艺参数,实现生物离心机工艺参数优化建模。结果表明,本算法预测误差和预测相对误差最高仅为0.2和0.05,均低于SVM算法、PSO-SVM和GSO-SVM三种算法,说明采用本算法的预测精度更高,可加快SVM算法计算速度,实现快速收敛和全局最优,更适用于离心机工艺参数优化和模型构建。

To improve the accuracy of process parameters by SVM.The PSO-SVM algorithm is improved by Gaussian kernel function RBF,mainly from the kernel function center calculation,kernel function width calculation and P S M function optimization S V O algorithm,thus obtaining the RBF-PSO-SVM method to obtain the optimal process parameters for biological centrifuge.The results show that the prediction error and relative error of this algorithm are only 0.2 and 0.05,which are lower than SVM algorithm,PSO-SVM and GSO-SVM.It shows that higher prediction accuracy of this algorithm can speed up SVM algorithm and achieve fast convergence and global optimization;it is more suitable for centrifugal process parameter optimization and model construction.