食品科学 ›› 2010, Vol. 31 ›› Issue (6): 205-207.doi: 10.7506/spkx1002-6630-201006048

• 分析检测 • 上一篇    下一篇

大豆油脂过氧化值的近红外光谱分析

王立琦1,张礼勇1 ,* ,朱秀超2   

  1. 1.哈尔滨理工大学测控技术与通信工程学院 2.哈尔滨商业大学计算机与信息工程学院
  • 收稿日期:2009-05-31 修回日期:2009-09-11 出版日期:2010-03-15 发布日期:2010-12-29
  • 通讯作者: 张礼勇1 E-mail:hlgzly@163.com

Near-infrared Spectroscopic Analysis of Peroxide Value of Soybean Oil

WANG Li-qi1,ZHANG Li-yong1,*,ZHU Xiu-chao2   

  1. (1. School of Measurement-Control Tech and Communications Engineering, Harbin University of Science and Technology, Harbin
    150086, China;2. School of Computer and Information Engineering, Harbin University of Commerce, Harbin 150028, China)
  • Received:2009-05-31 Revised:2009-09-11 Online:2010-03-15 Published:2010-12-29
  • Contact: ZHANG Li-yong1 E-mail:hlgzly@163.com

摘要:

首先采用碘量法测定油脂样品的过氧化值作为校正值,再使用近红外光谱仪采集油脂样品的光谱数据。在油脂过氧化值的定量分析中建立3 种校正模型,对比分析偏最小二乘回归(PLS)、多元线性回归(MLR)和主成分回归(PCR)的性能,确定最优模型为PLS。使用PLS 测定油脂过氧化值的性能结果为:校正集的相关系数0.916,预测集的相关系数0.922。

关键词: 近红外光谱, 大豆油脂, 过氧化值

Abstract:

In this study, peroxide values of 255 soybean oil samples measured by iodimetric assay were used for calibration and near-infrared spectral data of these samples were collected. In the quantitative analysis of peroxide value, three models were established using partial least squares regression (PLS-regression), multiple linear regression (MLR) and principal component regression (PCR). Comparative analysis indicated that among the established models, PLS-regression model exhibited the best reliability in predicting peroxide value of soybean oil, with correlation coefficients of calibration and validation sets of 0.916 and 0.922, respectively.

Key words: near-infrared spectroscopy, soybean oil, peroxide value

中图分类号: