首页 | 本学科首页   官方微博 | 高级检索  
     检索      

近红外光谱用于植物样品中水溶性氯离子含量的测定
引用本文:吴荣晖,邵学广.近红外光谱用于植物样品中水溶性氯离子含量的测定[J].光谱学与光谱分析,2006,26(4):617-619.
作者姓名:吴荣晖  邵学广
作者单位:中国科学技术大学化学系,安徽,合肥,230026
摘    要:基于离散小波变换(DWT)和最小二乘支持向量回归(LSSVR)方法, 建立了近红外光谱测定植物样品中水溶性氯离子的回归校正模型. 以烟草样品中水溶性氯离子含量的测定为研究对象, 首先采用DWT对近红外光谱进行数据压缩和背景扣除, 再用LSSVR建立氯离子的校正模型. 结果表明, 与偏最小二乘回归(PLSR)和传统的LSSVR方法相比, 作者所建模型的预测准确性具有一定优势.

关 键 词:离散小波变换  最小二乘支持向量回归  近红外光谱  水溶性氯离子
文章编号:1000-0593(2006)04-0617-03
收稿时间:2005-03-08
修稿时间:2005-06-18

Application of Near-Infrared Spectra in the Determination of Water Soluble Chloride Ion in Plant Samples
WU Rong-hui,SHAO Xue-guang.Application of Near-Infrared Spectra in the Determination of Water Soluble Chloride Ion in Plant Samples[J].Spectroscopy and Spectral Analysis,2006,26(4):617-619.
Authors:WU Rong-hui  SHAO Xue-guang
Institution:Department of Chemistry, University of Science and Technology of China, Hefei 230026, China
Abstract:A new method was proposed to extract relevant information from near-infrared(NIR) spectra for multivariate calibration of water soluble chloride ion in complex plant samples.The method is a combination of discrete wavelet transform(DWT) and least squares support vector regression(LSSVR).After data compression and background removal in NIR spectra by DWT,the LSSVR approach was used to build NIR spectra regression models on the retained wavelet coefficients.Compared with(partial) least square regression(PLSR) and LSSVR,the proposed method is superior both in calculation speed and prediction(accuracy).
Keywords:Discrete wavelet transform  Least squares support vector regression  Near-infrared spectra  Water soluble chloride
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号