摘要
核主元分析(KPCA)方法通过核变换将输入空间映射到高维特征空间,在特征空间进行主元分析。由于KPCA不适合大样本数据建模与分析,因此建模数据的选取非常重要,合理的数据样本可以简化运算,提高核主元分析的诊断准确度。文章提出一种优化数据样本的KPCA方法,利用相似度函数的方法实现样本优化,再建立核主元分析模型,提取数据特征信息,并将该方法应用到水环境监测的传感器故障诊断中,通过试验分析,验证了该方法的有效性。
The kernel principal component analysis （KPCA） method can map the input space into the high dimensional feature space via kernel function and then make the principal component analysis in the feature space. The selection of modeling data samples is very important, that＇ s because the KPCA method is not suitable for the analysis of a large number of data samples. The appropriate data samples can reduce the calculation and improve the diagnostic accuracy. The kernel principal component analysis based on sample optimization is proposed in this paper, which can optimize the data samples using the similarity function and then establish the KPCA model, the data characteristic information is extracted via the model. The method is applied to the fault diagnosis of sensors of water quality monitoring. The result shows the efficiency of this method.
出处
《中国环境监测》
CAS
CSCD
北大核心
2012年第2期92-96,共5页
Environmental Monitoring in China
基金
河海大学常州校区创新基金项目(XZX/09B002-02)
关键词
核主元分析
相似度函数
样本优化
水环境监测
传感器故障诊断
Kernel principal component analysis
Similarity function
Sample optimization
Water quality monitoring
Fault diagnosis
