摘要
本文采用近红外光谱技术采集样品的近红外光谱数据,光谱经一阶求导后,采用偏最小二乘法(PLS)建立花生油中棕榈油含量的定标模型,并用交互验证法对模型进行了验证。模型相关系数为0.9963,校正均方根(RMSEC)为0.937。该模型应用于实际样品的检测,结果令人满意。
Near-Infrared reflectance spectroscopic data have been collected and treated with chemometric methodology.After taking the first order derivative spectrum,a calibration model for determining the palm oil content in peanut oil samples was established by using partial least squares method（PLS）,and the cross validation in chemometrics was used for the model testing.The correlation coefficient was 0.9963and the root mean square error of calibration（RMSEC）was 0.937.This method has been successfully used in the analysis of actual samples with satisfactory results.
出处
《分析科学学报》
CAS
CSCD
北大核心
2010年第6期673-676,共4页
Journal of Analytical Science
基金
江西省科技支撑计划(No.2009JX01703)
关键词
近红外光谱
花生油
棕榈油
偏最小二乘法
Near infrared spectroscopy
Peanut oil
Palm oil
Partial least squares
