白鹏利,王钧,尹焕才,殷建,田晶晶,陈名利,高静.基于偏最小二乘留一交叉验证法的近红外光谱建模样品选择方法的研究[J].食品安全质量检测学报,2017,8(1):182-186
基于偏最小二乘留一交叉验证法的近红外光谱建模样品选择方法的研究
Study on the sample selection methods for building calibration model by near infrared spectroscopy based on partial least squares-leave one out-cross validation
投稿时间:2016-11-24  修订日期:2017-01-12
DOI:
中文关键词:  近红外光谱  偏最小二乘留一交叉验证法  样品挑选  定标集
英文关键词:near infrared spectroscopy  partial least squares-leave one out-cross validation  sample selection  calibration set
基金项目:国家高技术研究发展计划(863计划)项目(2015AA021106)
作者单位
白鹏利 中国科学院苏州生物医学工程技术研究所 
王钧 中国科学院苏州生物医学工程技术研究所 
尹焕才 中国科学院苏州生物医学工程技术研究所 
殷建 中国科学院苏州生物医学工程技术研究所 
田晶晶 中国科学院苏州生物医学工程技术研究所 
陈名利 中国科学院苏州生物医学工程技术研究所 
高静 中国科学院苏州生物医学工程技术研究所 
AuthorInstitution
BAI Peng-Li Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Science 
WANG Jun Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Science 
YIN Huan-Cai Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Science 
YIN Jian Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Science 
TIAN Jing-Jing Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Science 
CHEN Ming-Li Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Science 
GAO Jing Suzhou Institute of Biomedical Engineering and Technology, Chinese Academy of Science 
摘要点击次数: 1279
全文下载次数: 1809
中文摘要:
      目的 提出一种新的挑选定标集的方法-偏最小二乘留一交叉验证法。方法 以玉米为例, 通过对玉米中水分含量的实际建模与外部验证, 根据主成分数、相关系数、预测均方根差以及相对分析误差(ratio of performance to standard deviate, RPD)等因素, 综合比较4种定标集挑选方法的优缺点。结果 偏最小二乘留一交叉验证法结合样品和光谱性质, 在保持原始样品覆盖范围的基础上, 挑选出的定标集所建立的模型具有较低的模型复杂程度、较高的验证相关系数以及较高的RPD值。结论 该方法既克服了随机挑选法存在的样品代表性不足的风险, 同时也避免了含量梯度法和计算机识别法只考虑样品或者光谱的单一性质的不足, 同时该方法具有操作简单、易于推广等优点, 为食品安全检测提供了一种新的筛选样品的方法。
英文摘要:
      Objective To establish a new method of selecting calibration set-partial least squares-leave one out-cross validation (PLS-LOO-CV). Methods According to the values of principal component number, correlation coefficient (R), root mean square error of prediction (RMSEP), and ratio of performance to standard deviate (RPD), the 4 selection methods of calibration sets were compared by modeling and validation of water content in rice. Results By PLS-LOO-CV combined with samples and spectral properties, the selected sets of the model had lower model complexity, higher correlation coefficient and higher RPD value on the basis of maintaining the original sample coverage. Conclusion The established method can avoid the risk of lacking of representation caused by random method as well as the efficiency of considering single property of sample or spectrum using content grads method or computer recognition method. At the same time, the PLS-LOO-CV method has the advantages of simple operation and easy popularization, which provides a novel method of screening food samples for food safety inspection.
查看全文  查看/发表评论  下载PDF阅读器