摘要
目的在微量物证领域,元素分析是进行泥土物证检验的有效手段之一,X射线荧光光谱法(XRF)具有分析速度快、检验无损、可进行无标样分析等特点在泥土物证的比对检验中应用广泛。但是针对两个样品间多维泥土元素检验数据的相似性评价一直是困扰微量物证工作者的一个难题。本文旨在对比欧氏距离和Hotelling T2这两种统计学方法在评价样品检验数据差异性方面的准确性。方法采用XRF法对泥土样品进行10种元素定量分析,采用欧氏距离和Hotelling T2两种方法对检验数据进行统计处理,比较两种方法在评价两个样品间相似性时的差异。结果欧氏距离和Hotelling T2法在处理20个(190组)不同来源的泥土样品之间的两两比对时,不可区分率分别为4.21%和7.37%。结论欧氏距离和Hotelling T2两种多参数数据比对检验法均可用于泥土物证XRF定量检验数据间相似性评价,两种方法各有特点,在实际应用中要注意其使用条件和前提,才能保证统计结果的准确性。
Objective XRF(X-ray fl uorescence spectrometry)possesses many advantages for soil elemental quantitative analysis,yet leaving difficulties to evaluate similarity/difference between sampling soils with the received analysis data.Thus,two methods,Euclidean Distance and Hotelling T2-both of which are the statistical approaches to compare intersample similarity/difference,were adopted to evaluate their accuracy in comparison of data similarity/difference.Methods XRF was used to analyze the concentrations of 10 elements in 20 sampling soils collected from different locations,having the relevant data obtained.The methods of Euclidean Distance and Hotelling T2 were carried out to assess the difference between the data from sampling soils.Results With the 20 soil samples and their resultant 190 pairs of comparison,Euclidean Distance approach gave its indistinguishable rate as 4.21%,and Hotelling T27.37%.Conclusions Both Euclidean Distance and Hotelling T2 are applicable to statistically evaluate inter-soil samples.Yet,proper selection and execution of the statistical processing should be based on the prerequisites of the methods,thereby leading to acceptable results for sample comparison.
作者
郭洪玲
王萍
朱军
胡灿
梅宏成
郑继利
GUO Hongling;WANG Ping;ZHU Jun;HU Can;MEI Hongcheng;ZHENG Jili(Institute of Forensic Science,Ministry of Public Security,Beijing 100038,China)
出处
《刑事技术》
2020年第6期568-571,I0004-I0006,共7页
Forensic Science and Technology
基金
“十三五”国家重点研发计划项目(2017YFC0803803)。