|
|
此文章由 Hetbert 原创或转贴，不代表本站立场和观点，版权归 oursteps.com.au 和作者 Hetbert 所有！转贴必须注明作者、出处和本声明，并保持内容完整
本帖最后由 Hetbert 于 2018-12-13 15:50 编辑
hornsay 发表于 2018-12-12 19:47 
几点想法：
1. 怎么样的概率算是有意义的？True Allele有没有这方面的解释？或者说他们的概率是通用的？ ...
这个数字是有标准的，但是TrueAllele用的是倍数，TrueAllele的数字总是多了好多位数，其它软件数值是几万的时候，TrueAllele可能就是几百万到几十亿了，所以检方特别喜欢用TrueAllele，成功率特别高，陪审团看到一两万的数字，将信将疑，看到一二十亿，甚至18个0，15个0的这种理解不能的天文数字，就坚信不疑了。
The results for Stain 91 (Item 550) were processed on the assumption that it contained three, four or five contributors. Dr Perlin expressed the following opinions based upon the results of the TrueAllele processing:
(a) a match between a contributor to Stain 91 and Yun Bin Lin is 4,410 times more probable than a coincidental match to an unrelated Asian person; （这个值相当低，林云缤DNA到底有没有是不确定的.就是所说的Neutral，如果假设是四人DNA混合可以有，假设是三人DNA排除也可以。 ）
统计学上说的可能，不可能。和一般人理解的“可能”，“不可能”，有很大差别。
我举个例子：在数轴上随机点一个点，这个点是有理数的可能性有多大，统计学说可能性是0。
检方自己的专家证人也对法官如是说：
Q: Will you tell me, Mr Walton, where there was the same number of alleles in the two samples, that is, the reference samples, as combined and stain 550, what, if anything, can you say about the issue of peak heights, namely the intensity of the presentation of those alleles?
A: Yes. In my conclusion of this particular profile, the first thing ‑ when we do the analytical process or looking at a profile, the first thing we do is we try and determine the number of people that we think are present. We then look at the reference samples and whether we can exclude or include particular individuals.
If we can exclude, that is fine; if we can't, and there's the possibility they are included, we would then try and provide a statistical analysis of how likely that particular profile could be in the sample. The reason for this being that, if there is a profile that matches a part of the sample, then that means either that the DNA comes from that person or it comes from someone else and then the statistic gives the chance of how that could occur, how likely that is to occur.
检方所谓的血渍91有至少四名受害者的DNA的说法就充分立功了统计学里的“可能”和人们一般认知的“可能”的偏差。
https://www.caselaw.nsw.gov.au/decision/58aa87f9e4b058596cba431e
(b) a match between a contributor to Stain 91 and Yun Li Lin is 27.1 times more probable than a coincidental match to an unrelated Asian person; （林云莉的DNA应该是没有的，因为两个儿子有她的DNA，所以一定不可能排除。）
(c) a match between a contributor to Stain 91 and Min Lin is 379,000 times more probable than a coincidental match to an unrelated Asian person; (林暋的DNA在阀值以上，应该是有的。)
(d) a match between a contributor to Stain 91 and Henry Lin is 1.33 billion times more probable than a coincidental match to an unrelated Asian person; and （应该有Henry的DNA）
(e) a match between a contributor to Stain 91 and Terry Lin is 1.03 quintillion times more probable than a coincidental match to an unrelated Asian person. （一定有Terry的DNA） |
评分
-
查看全部评分
|
2024秋 第一次日本的三人行 从古韵幽幽到繁华流光(完) (2024-12-2) yeqingyezhunan 
西澳outback (2011-5-30) cigmm 
楼主:satellite0
发表于 2018-12-12 18:30
,孜孜不倦长辩到天明啊，业界良心天团。
