期刊文献+

基于多维光谱的多组分混合气体浓度支持向量机算法 被引量:3

Algorithm of Mixed Gas Concentration Analysis Based on Support Vector Machine and Multidimensional Spectrum
下载PDF
导出
摘要 针对多组分混合气体特征吸收光谱重叠严重的难点,提出一种新的基于多维光谱的多组分混合气体浓度支持向量机定量分析算法。算法采用核函数变换的方式,将重叠严重和非线性的光谱数据进行高维空间变换后再计算各组分气体浓度。将算法用于天然气各组分气体浓度的定量分析中,实验结果表明,算法的检验准确率和效率均显著优于其它多组分混合气体浓度定量分析算法。 According to the difficulty of serious overlap of mixed gas feature absorption spectrum, a novel algorithm of mixed gas concentration analysis based on support vector machine and multidimensional spectrum is proposed. The transformation of kernel function is used to solve the overlapped mixed gas feature absorption spectrum in high dimension space. After transformation, the mixed gas feature absorption spectrum is used to analyze the concentration of mixed gas. When the algorithm is applied in the quantitative analysis of natural gas components concentration, the experimental results show that the accuracy and efficiency of this algorithm are noticeably superior to other algorithms of its kind.
作者 白鹏 刘君华
出处 《化工自动化及仪表》 EI CAS 2005年第5期43-47,共5页 Control and Instruments in Chemical Industry
基金 国家自然科学基金资助项目(60276037)
关键词 支持向量机 回归 定量分析 混合气体 多维光谱 support vector machine regression quantitative analysis mixed gas multi dimensional spectrum
  • 相关文献

参考文献12

二级参考文献39

共引文献2438

同被引文献36

引证文献3

二级引证文献41

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部