摘要
研究近红外光谱技术对食用醋品牌进行快速无损溯源。收集市场上保宁、东湖、恒顺、镇江4个品种共152份具有代表性的食用醋样品,采集它们的近红外光谱数据,对原始光谱数据进行多元散射校正(multip,licative scatter corrertion,MSC)预处理,对预处理后的光谱数据利用主成分分析法(principal component analysis,PCA)进行聚类分析,根据主成分的累计贡献率选取主成分数,然后用支持向量机(support vector machine,SVM)建立预测模型,选取合适的SVM核函数,并利用粒子群优化算法(particle swarm optimization,PSO)优化模型参数。结果表明,近红外光谱技术结合支持向量机对食用醋品牌分类正确率可达100%。
Presented a fast and non-destructive method for the dis crimination of vinegar brands by nearinfrared spectroscopy technology. One hundred and fifty two representative samples of vinegar including Bao Ning, East Lake, Heng Shun, Zhenjiang were collected from market. Multiplicative Scatter Correction （MSC） was used to handle the original near infrared spectrum （NIR） data and Principal Component Analysis （PCA） was used to process the spectral data after pretreatment according to the accumulative contribution rate of principal components to select principal components. Support Vector Machine （SVM） was then applied to build the brand traceabilitymodel with proper kernel function. Particle Swarm Optimization was applied to optimize the parameters of the model. The experiments indicated that the method combing near infrared spectroscopy with support vector machine could classify the vinegar brand with 100% accuracy.
出处
《食品与机械》
CSCD
北大核心
2016年第1期38-40,50,共4页
Food and Machinery
基金
上海市自然科学基金(编号:14ZR1419200)
关键词
近红外光谱
主成分分析法
支持向量机
粒子群优化算法
品牌溯源
near infrared spectroscopy
principal component analysis
support vector machine
particle swarm optimization
brand tracea bility
