基于粒子群优化支持向量机的瑞芬太尼血药浓度预测模型

Remifentanil Blood Concentration Forecast Model based on Support Vector Machine with Particle Swarm Optimization

目的建立基于粒子群优化算法的瑞芬太尼血药浓度支持向量和模型。方法本实验采用粒子群算法(particle swarm optimization,PSO)优化支持向量机(support vector machine,SVM)算法,建立粒子群优化支持向量机(PSO-SVM)瑞芬太尼血药浓度预测模型。该模型能从较少的采样数据中准确捕捉血药浓度和时间、病人体征、给药方案之间的非线性关系。结果粒子群优化支持向量机的平均误差为-1.07%,非线性混合效应模型(nonlinear mixed effects modeling,NONMEM)为-2.24%,粒子群优化支持向量机网络的绝对平均误差9.09%,非线性混合效应模型为19.92%。结论粒子群优化支持向量机模型能迅速,稳定预测瑞芬太尼血药浓度,且准确度高,误差较小。该方法原理简单,实现便捷,运算速度快,适用于半衰期较短的麻醉速效药等多房室结构药物的群体药代药效学研究和分析。

OBJECTIVE To develop a SVM model which is constructed by using particle swarm optimization to a predict the plasma concentration of remifentail. METHODS This research establishes a PSO-SVM model which is constructed by using particle swarm optimization to a predict the plasma concentration of remifentanil. The model was capable of capturing the nonlinear relationship among plasma concentration, time, and the patient＇s signs exactly. RESULTS The average error of PSO-SVM is - 1.07% , while that of NONMEM is - 2.24%. The absolute average error of PSO-SVM is 9.09% , while that of NONMEM is 19.92%. CONCLUSION Experimental results indicate that PSO-SVM model could predict the plasma concentration of remifentanil rapidly and stably, with high accuracy and low error. For the characteristic of simple principle and fast computing speed, this method is suitable to data analysis of short-acting anesthesia drug population pharmacokinetics and pharmacodynam- ics,