通过寄递渠道传播的含有非法添加成分的减肥药物是公安机关侦办食药环犯罪的重点打击对象。为快速鉴别此类药物，本研究采用分子光谱分析技术，对含艾司唑仑、西地那非、西布曲明、氟硝西泮、唑吡坦等5种精神管制类药物成分的“减肥药”进行检验，获取了145组光谱数据。采用主成分分析提取主成分因子对数据降维。基于所提取的20维数据，建立FDA模型、KNN模型、SVM模型并进行对比。在模型1中构建3个Fisher判别函数对5类样品进行判别，准确度达到100%；在模型2中K值的变化会影响分类器精度，通过对K值的调整能够快速对5类样品进行分类，准确率达到100%；在模型3中选用RBF核函数，分别对比唑吡坦与其他4类减肥药物分类效果，准确率均达到100%。通过实验中的数据集对唑吡坦不同品牌的样本进行识别和对实际案件进行分析，对公安机关侦办此类案件具有一定参考。

Slimming medicines containing illegal additives, often distributed through mail and other delivery channels, are primary targets for public security organizations to curb food, drug, and environmental crimes. To enable rapid identification of such drugs, this study adopted molecular spectral analysis technology to examine slimming medication containing estazolam, sildenafil, sibutramine, flurazepam, and zolpidem; 145 groups of spectral data were obtained. Principal component analysis was utilized to extract the principal component factors and reduce data dimension. Based on the extracted 20-dimensional data, the Fisher discriminant analysis (FDA), K-nearest neighbor (KNN), and support vector machine models were established for comparison. In model 1, three Fisher discriminant functions were constructed to discriminate five types of samples, and the accuracy reached 100%. In model 2, the change of K value will affect the accuracy of the classifier. Through the adjustment of K value, 5 kinds of samples can be classified quickly, with an accuracy rate of 100%. In model 3, RBF kernel function was used to compare the classification effect of zolpidem and other four kinds of slimming drugs, and the accuracy rate reached 100%. Through the dataset in the experiment, the samples of different brands of zolpidem are identified and the actual cases are analyzed, which has a certain reference for the public security organs to investigate such cases.