基于支持向量机(SVM)在处理小样本、高维数及泛化性能强等方面的优势,提出了基于主元分析和支持向量机(PCA—SVM)对过程进行监控的方法。文中先利用主元分析方法进行特征数据提取,得到降维的主元特征向量,去除了高维样本变量相关性。然后分析各状态T^2统计、SPE统计量的变化趋势,对实际生产状况进行监控,最后利用SVM与最近邻法相结合的策略对特征向量进行分类识别。试验结果证实了提出算法的有效性。
Based on the high performance of support vector machine (SVM) in tackling small sample size, high dimension and its good generalization, a process monitoring method hased on principal component analysis (PCA) and SVM is proposed. Firstly, the PCA approach is adopted to extract the feature and reduce the dimension of data by getting rid of the correlation among them, and then it is applied to statistical process control of the imperial smelting furnace (ISF), with the change trend of expectations of T^2 and SPE statistics of the data, the ISF manufacture states are rested. Finally, the SVM combined with the nearest neighbor method is used for classification. The experiment result shows that the method is effective.