基于似然比方法的法庭证据评估  

Evaluation of Evidence Based on Likelihood Ratio Method

作　　者：董锋 罗亚平 赵雅彬 李康 张伟方 崔恒建 DONG Feng;LUO Ya-ping;ZHAO Ya-bin;LI Kang;ZHANG Wei-fang;CUI Heng-jian(School of Forensic Science,People's Public Security University of China,Beijing 100038,China;Graduate School,People's Public Security University of China,Bejjing 100038,China;Shanghai Key Laboratory of Crime Scene Evidence,Shanghai Research Institute of Criminal Science and Technology,Shanghai 200080,China;School of Mathematical Sciences,Capital Normal University,Beijing 100048,China)

机构地区：[1]中国人民公安大学刑事科学技术学院,北京100038 [2]中国人民公安大学研究生院,北京100038 [3]上海市现场物证重点实验室,上海200080 [4]首都师范大学数学科学学院,北京100048

出　　处：《数理统计与管理》2021年第1期15-25,共11页Journal of Applied Statistics and Management

基　　金：上海市刑事科学技术研究院物证重点实验室开放课题(2018XCWZK18);国家自然科学基金项目(11971324,11471223);北京市科技创新平台建设项目;首都师范大学交叉研究院项目以及生物信息与统计交叉学科项目。

摘　　要：在法庭科学领域,如何客观表述证据证明力大小给检验人员带来了一些困难。针对这一问题,本文介绍了两种似然比方法,利用概率密度函数计算支持相关假设的敏感性与特异性。通过枪弹痕迹鉴定的实证分析,结合评分法似然比的计算及评价方式,综合比较了边缘似然比方法、联合似然比方法以及支持向量机分类方法,结果表明联合似然比方法能够获得较低的错误率。本文深入探索了似然比方法在法庭科学领域中的适用性与可行性,旨在进一步推动我国法庭科学证据量化和科学化评估进程。There are some questions corresponding to the objective evaluation about evidence in the field of forensic science.To deal with these problems,this paper introduces two methods of likelihood ratio which use the probability density function to calculate the sensitivity and specificity supporting the relative proposition.An example of firearm examination is proposed and researched with the calculation and evaluation of likelihood ratio method,the false positive rates and false negative rates were compared among the marginal likelihood ratio method,the joint likelihood ratio method and the supported vector machine.The results showed that the joint likelihood ratio method was better than the other method.This paper investigated the application and feasibility of likelihood ratio method in the field of forensic science.This work acts as a move toward an objective and transparent evaluation process for the evaluation of evidence.

关 键 词：似然比 法庭科学 证据评估 

分 类 号：DF794.1[政治法律—诉讼法学] O212[政治法律—法学]