KNN算法在动物油鉴别区分中的应用研究
发布时间：2019-05-15 00:28
【摘要】：气相色谱-质谱联用法(GC-MS)因其分离效率高、分析速度快、灵敏度高、检测线低等特点,被广泛应用于油脂分析鉴别领域。但与矿物油的鉴别相比,动物油类之间的主成分种类相近且含量集中,单纯通过GC-MS进行鉴别分析具有局限性,因此,区分常见动物油一直是司法鉴定领域中的难题。本文尝试运用GC/MS分析人油和5种常见动物油,通过对峰面积归一化法得出每个样品脂肪酸相对百分含量,结合KNN算法(K Nearest Neighbors,KNN)对人油与常见动物油进行建模区分。本实验以每个动物油脂样本中的6个主要脂肪酸相对含量(C14:0、C16:0、C16:1、C18:0、C18:1、C18:2)作为变量值,运用训练样本即为测试样本的方法进行交互验证,发现当k值等于3或4时,测试样本出错率最低,区分效果良好,人油测试样本分类准确率达到100%,并考察了6种脂肪酸相对含量作为变量的区分贡献值,结果C14:0区分贡献值最大。此方法相对于传统分析手段而言简单易行,提高了鉴别分析的效率和精度,尽管实验样本种类有限,但实验方法具有普遍意义。本文为动物油区分的进一步深入研究提供了一种新的思路和参考。
[Abstract]:Gas chromatography-mass spectrometry (GC-MS) has been widely used in the field of oil analysis and identification because of its high separation efficiency, fast analysis speed, high sensitivity and low detection line. However, compared with the identification of mineral oil, the principal components of animal oil are similar and concentrated, and the identification analysis by GC-MS alone has limitations. Therefore, distinguishing common animal oil has always been a difficult problem in the field of judicial identification. In this paper, GC/MS is used to analyze human oil and five kinds of common animal oil, and the relative fatty acid content of each sample is obtained by normalizing the peak area. Combined with KNN algorithm (K Nearest Neighbors,KNN), the human oil and common animal oil are modeled and distinguished. In this experiment, the relative contents of six major fatty acids in each animal oil sample (C14, C16, C18, C1, c18, c18) were taken as variable values, and the relative contents of 6 major fatty acids in each animal oil sample (C14, C16, C18, C1, c18) were used as variable values. When the k value is 3 or 4, the error rate of the test sample is the lowest, the distinguishing effect is good, and the classification accuracy of the human oil test sample is 100%. The differential contribution values of the relative contents of six fatty acids as variables were investigated, and the results showed that the differential contribution value of C14 鈮，
本文编号：2477168
[Abstract]:Gas chromatography-mass spectrometry (GC-MS) has been widely used in the field of oil analysis and identification because of its high separation efficiency, fast analysis speed, high sensitivity and low detection line. However, compared with the identification of mineral oil, the principal components of animal oil are similar and concentrated, and the identification analysis by GC-MS alone has limitations. Therefore, distinguishing common animal oil has always been a difficult problem in the field of judicial identification. In this paper, GC/MS is used to analyze human oil and five kinds of common animal oil, and the relative fatty acid content of each sample is obtained by normalizing the peak area. Combined with KNN algorithm (K Nearest Neighbors,KNN), the human oil and common animal oil are modeled and distinguished. In this experiment, the relative contents of six major fatty acids in each animal oil sample (C14, C16, C18, C1, c18, c18) were taken as variable values, and the relative contents of 6 major fatty acids in each animal oil sample (C14, C16, C18, C1, c18) were used as variable values. When the k value is 3 or 4, the error rate of the test sample is the lowest, the distinguishing effect is good, and the classification accuracy of the human oil test sample is 100%. The differential contribution values of the relative contents of six fatty acids as variables were investigated, and the results showed that the differential contribution value of C14 鈮，
本文编号：2477168
本文链接：https://www.wllwen.com/falvlunwen/fanzuizhian/2477168.html
