首页    期刊浏览 2024年11月25日 星期一
登录注册

文章基本信息

  • 标题:階層型クラスター分析を用いた大麻異同識別法の検討
  • 本地全文:下载
  • 作者:山室 匡史 ; 桑山 健次 ; 辻川 健治
  • 期刊名称:日本法科学技術学会誌
  • 印刷版ISSN:1880-1323
  • 电子版ISSN:1881-4689
  • 出版年度:2016
  • 卷号:21
  • 期号:2
  • 页码:109-124
  • DOI:10.3408/jafst.709
  • 出版社:Japanese Association of Forensic Science and Technology
  • 摘要:A method to distinguish the presence or absence of relationship between two cannabis seizures is required for forensic investigations. In this study, we focused on the chemical profile of the cannabis. We developed a method to discriminate whether the two cannabis seizures derived from a single organism or different ones (“individual” discrimination) and whether the two seizures had been seized in the same criminal investigation or not (“case” discrimination). Discrimination ability of these methods were evaluated statistically. For 21 cannabis seizures, gas chromatography/time-of-flight mass spectrometry analysis were conducted to obtain the comprehensive chemical profile. The contents of not only cannabinoids and terpenes but almost all detected compounds were used as variables to perform multivariate analysis. Using hierarchical cluster analysis, most of the cannabis seized in an identical case formed one group in the dendrogram. The cannabis cultivated outdoors and indoors belonged to different groups. 441 hierarchical cluster analysis were performed with a combination of all pair of cannabis seizures to calculate the Euclidean distance between specimens based on definite rules. Two thresholds were arbitrarily set for the distance to perform two types of cannabis discrimination. The equal error rate and the area under the receiver operating characteristic curve of “individual” discrimination were 0.024 and 0.997 and those of “case” discrimination were 0.220 and 0.867. By selecting an appropriate threshold, “individual” discrimination was achieved with sensitivity of 0.810, specificity of 0.998 and accuracy rate of 0.989. Similarly, “case” discrimination was with sensitivity of 0.627, specificity of 0.971 and accuracy rate of 0.925. These results suggested that the present method by using hierarchical cluster analysis of chemical profile would be effective for cannabis discrimination in forensic investigations.
  • 关键词:Cannabis;chemical profile;GC/TOF-MS;hierarchical cluster analysis
国家哲学社会科学文献中心版权所有