o**********e
发帖数: 18403 | 1 【 以下文字转载自 JobHunting 讨论区 】
发信人: onetiemyshoe (onetiemyshoe), 信区: JobHunting
标 题: 发现一数据造假的老印 (转载)
发信站: BBS 未名空间站 (Thu Aug 13 11:02:56 2015, 美东)
发信人: hopeusw (云子), 信区: Faculty
标 题: 发现一数据造假的老印
发信站: BBS 未名空间站 (Tue Aug 11 19:04:30 2015, 美东)
本来是帮一个杂志审稿,这人开头就说他的计算机模型的准确率是100%,因为他是做动物
数据模拟的,本身实验数据的准确率能有80%就不错了.
我就查了查他的数据,非常简单的造假,就是如果预测的结果不好,这个数据他就不用了,
理由是处理不了,然后有的是预测结果是2.3,然后就把原始数据从2.8给改成2.3,如此种
种.
我又翻翻这位最近的PAPER,光今年就好几篇类似的,都是号称100%的准确率,这么好的结
果,挑的都是3分左右的杂志发,我刚给里面最好的一个杂志的主编写了封信,看看他怎么
说.
另外我不知道好多审稿子的有没有常识,如果数据本身的准确率就80%左右,能做出来100
%的准确率的计算机模型,是能把噪声都给OPTIMIZE了吗?
Another famous/salacious/highly representative case of academic fraud also
resume fraud:
http://www.nbcnews.com/news/us-news/ivory-tower-phony-sex-lies-
"Ivory Tower Phony? Sex, Lies and Fraud Alleged in W. Va. University Case"
He seemed like the Doogie Howser of India, able to crack the country’s best
medical school, and work there as a 21-year-old doctor. Anoop Shankar later
claimed to add a Ph.D. in epidemiology and treat patients even as he
researched population-wide diseases. He won a “genius” visa to America,
shared millions in grants, and boasted of membership in the prestigious
Royal College of Physicians.
In 2012 West Virginia University hand-picked this international star to help
heal one of the country’s sickest states. At just 37, Shankar was
nominated to the first endowed position in a new School of Public Health,
backed by a million dollars in public funds. As chair of the epidemiology
department, he was also poised to help the university spend tens of millions
of additional tax dollars. “This is about improving healthcare and
improving lives,” said university president Jim Clements, announcing a
federal grant for health sciences. “We could not be more proud.” | a***e
发帖数: 27968 | 2 靠，咋没人认真挖挖CTO&CFO of USA
【在 o**********e 的大作中提到】
: 【 以下文字转载自 JobHunting 讨论区 】
: 发信人: onetiemyshoe (onetiemyshoe), 信区: JobHunting
: 标 题: 发现一数据造假的老印 (转载)
: 发信站: BBS 未名空间站 (Thu Aug 13 11:02:56 2015, 美东)
: 发信人: hopeusw (云子), 信区: Faculty
: 标 题: 发现一数据造假的老印
: 发信站: BBS 未名空间站 (Tue Aug 11 19:04:30 2015, 美东)
: 本来是帮一个杂志审稿,这人开头就说他的计算机模型的准确率是100%,因为他是做动物
: 数据模拟的,本身实验数据的准确率能有80%就不错了.
: 我就查了查他的数据,非常简单的造假,就是如果预测的结果不好,这个数据他就不用了,
| o**********e
发帖数: 18403 | 3
this one?
http://www.dvorak.org/blog/2009/08/12/special-report-is-us-chie
which cfo?
【在 a***e 的大作中提到】
: 靠，咋没人认真挖挖CTO&CFO of USA
| o**********e
发帖数: 18403 | 4 【 以下文字转载自 JobHunting 讨论区 】
发信人: onetiemyshoe (onetiemyshoe), 信区: JobHunting
标 题: 发现一数据造假的老印 (转载)
发信站: BBS 未名空间站 (Thu Aug 13 11:02:56 2015, 美东)
发信人: hopeusw (云子), 信区: Faculty
标 题: 发现一数据造假的老印
发信站: BBS 未名空间站 (Tue Aug 11 19:04:30 2015, 美东)
本来是帮一个杂志审稿,这人开头就说他的计算机模型的准确率是100%,因为他是做动物
数据模拟的,本身实验数据的准确率能有80%就不错了.
我就查了查他的数据,非常简单的造假,就是如果预测的结果不好,这个数据他就不用了,
理由是处理不了,然后有的是预测结果是2.3,然后就把原始数据从2.8给改成2.3,如此种
种.
我又翻翻这位最近的PAPER,光今年就好几篇类似的,都是号称100%的准确率,这么好的结
果,挑的都是3分左右的杂志发,我刚给里面最好的一个杂志的主编写了封信,看看他怎么
说.
另外我不知道好多审稿子的有没有常识,如果数据本身的准确率就80%左右,能做出来100
%的准确率的计算机模型,是能把噪声都给OPTIMIZE了吗?
Another famous/salacious/highly representative case of academic fraud also
resume fraud:
http://www.nbcnews.com/news/us-news/ivory-tower-phony-sex-lies-
"Ivory Tower Phony? Sex, Lies and Fraud Alleged in W. Va. University Case"
He seemed like the Doogie Howser of India, able to crack the country’s best
medical school, and work there as a 21-year-old doctor. Anoop Shankar later
claimed to add a Ph.D. in epidemiology and treat patients even as he
researched population-wide diseases. He won a “genius” visa to America,
shared millions in grants, and boasted of membership in the prestigious
Royal College of Physicians.
In 2012 West Virginia University hand-picked this international star to help
heal one of the country’s sickest states. At just 37, Shankar was
nominated to the first endowed position in a new School of Public Health,
backed by a million dollars in public funds. As chair of the epidemiology
department, he was also poised to help the university spend tens of millions
of additional tax dollars. “This is about improving healthcare and
improving lives,” said university president Jim Clements, announcing a
federal grant for health sciences. “We could not be more proud.” | a***e
发帖数: 27968 | 5 靠，咋没人认真挖挖CTO&CFO of USA
【在 o**********e 的大作中提到】
: 【 以下文字转载自 JobHunting 讨论区 】
: 发信人: onetiemyshoe (onetiemyshoe), 信区: JobHunting
: 标 题: 发现一数据造假的老印 (转载)
: 发信站: BBS 未名空间站 (Thu Aug 13 11:02:56 2015, 美东)
: 发信人: hopeusw (云子), 信区: Faculty
: 标 题: 发现一数据造假的老印
: 发信站: BBS 未名空间站 (Tue Aug 11 19:04:30 2015, 美东)
: 本来是帮一个杂志审稿,这人开头就说他的计算机模型的准确率是100%,因为他是做动物
: 数据模拟的,本身实验数据的准确率能有80%就不错了.
: 我就查了查他的数据,非常简单的造假,就是如果预测的结果不好,这个数据他就不用了,
| o**********e
发帖数: 18403 | 6
this one?
http://www.dvorak.org/blog/2009/08/12/special-report-is-us-chie
which cfo?
【在 a***e 的大作中提到】
: 靠，咋没人认真挖挖CTO&CFO of USA
| o**********e
发帖数: 18403 | |
|
