基于红外光谱和最小二乘支持向量机建立掺杂牛奶与纯牛奶的判别模型。分别配置含有葡萄糖牛奶(0.01~0.3 gL-1)和三聚氰胺牛奶(0.01~0.3 gL-1)样品各36个, 采集纯牛奶及掺杂牛奶样品的红外光谱。采用最小二乘支持向量机分别建立掺杂葡萄糖、掺杂三聚氰胺、两种掺杂牛奶与纯牛奶的判别模型, 并利用这些模型对未知样品进行判别, 其判别正确率都为95.8%。研究结果表明: 与线性的偏最小二乘判别建模方法相比, 最小二乘支持向量机方法具有更强的预测能力。

The discrimination models for adulterated milk were established using infrared spectroscopy and least squares support vector machines (LS-SVM). 36 glucose-tainted milk(0.01~0.3 gL-1)and 36 melamine-tainted milk(0.01~0.3 gL-1)were prepared respectively. Then the infrared absorption spectra of all samples were measured. The LS-SVM models for glucose-tainted milk, melamine-tainted milk and two kinds adulterated milk were built. The classification accuracy rates for unknown samples all were 95.8%. The results showed that comparing with PLS-DA method; LS-SVM method had better prediction ability for complex milk system.