摘要
提出了一种将核主成分分析(KPCA)和最小二乘支持向量机(LSSVM)相结合的软测量建模方法。核主成分分析能够对样本数据进行特征提取,消除数据的相关性。本文利用KPCA提取主元,降低样本的维数;然后利用最小二乘支持向量机进行建模,不仅降低了模型的复杂性,而且提高了模型的泛化能力。用该方法建立柴油凝点的软测量模型,和其他4种方法比较,结果表明基于KPCA-LSSVM方法建立的软测量模型有较好的预测效果和泛化能力,是一种有效的数据建模方法。
A kind of soft sensing is proposed by combining kernel principal component analysis （KPCA） with least square support vector machine （LSSVM）. The KPCA has excellent performance of feature extraction and can eliminate the correlation of the input. The KPCA is used to choose the principal component and reduce dimensions of sample. Then LSSVM is applied to proceed regression modelling, which can not only reduce the complexity of modeling hut also improve the generalization ability. The proposed method is used to build soft sensing of diesel oil solidifying point. Compared with other four models, the result shows that KPCA-LSSVM approach has a better prediction and generalization, and is an effective data modeling.
出处
《化工学报》
EI
CAS
CSCD
北大核心
2011年第10期2813-2817,共5页
CIESC Journal
基金
国家高技术研究发展计划项目(2007AA04Z193)~~
