摘要
由于酒精发酵中pH过程的复杂性和强非线性,建立能够描述其动态和静态特性的准确的数学模型是比较困难的。支持向量机(SVM)是近年来发展起来的一种基于统计学习理论,采用结构风险最小化的新型学习机器。将支持向量机应用到酒精发酵生产过程pH值的建模中,实验和仿真结果表明支持向量机模型是有效的,且具有良好的泛化性能,能够较好的解决酒精发酵过程中pH值的建模问题。最后进一步分析了参数选择对支持向量机模型性能的影响。
Due to the complexity and highly non-linearity of the pH variable inherent in the process of alcohol fermentation, it is rather difficult to derive an exact mathematical model capable of describing both the dynamic and static characteristics. The support vector machine (SVM) is a novel modeling method based on the statistical learning theory, which employs the principle of the structural minimization. In this paper, a modeling scheme based on the SVM is proposed for the pH variable in an alcohol fermentation process. Both simulations and experimental results show the effectiveness of the proposed approach based SVM. The obtained model shows good performance in model validation. On the other hand, the effect of the selection of the parameters for SVM on the performance of the modeling procedure is also discussed.
出处
《微计算机信息》
北大核心
2008年第28期313-314,298,共3页
Control & Automation
基金
国家自然科学基金资助(60572055)项目名称:含有非平滑非线性的三明治动态系统辨识与预测
关键词
PH
酒精发酵
支持向量机
模型
pH
alcohol fermentation
support vector machine
model