Comments on: Classic Neyman-Pearson approach demo http://lixiaoxu.lxxm.com/classic-neyman-pearson-approach-demo/ Teaching notes of szpku dot lixiaoxu at gmail dot com Fri, 30 Jul 2010 07:54:24 +0000 http://wordpress.org/?v=2.7 hourly 1 By: lixiaoxu http://lixiaoxu.lxxm.com/classic-neyman-pearson-approach-demo/comment-page-1/#comment-74 lixiaoxu Sun, 02 Dec 2007 12:15:36 +0000 http://lixiaoxu.lxxm.com/2007/11/30/classic-neyman-pearson-approach-demo/#comment-74 ##If you guess the point has 22% probability to fall in the H_1 region, you will update the probability to 82% after you know it fall in the light gray region with a significant exp report. ##to demo Prosecutor's fallacy on wiki p_prior=10e-7;##if the defendant is not different From any one of the 10 million population except the DNA test result p_alpha=0.0001;##"it is testified that the probability that two DNA profiles match by chance is only 1 in 10,000" p_power=1; ##Let's assume the power is extreme ## (h_1_s=p_power*p_prior/(p_alpha*(1-p_prior)+p_power*p_prior)); ## the Procecutor mistook the 0.99% target percent as 99.99%. ##------------------------------- ##to demo the Defendant's fallacy p_prior=10e-2;##"The video (evidence) suggests a 99% (rather than 1-10e-7) chance that the defendant is innocent", except the DNA test result p_alpha=0.0001;##as the former p_power=1; ##as the former ## (h_1_s=p_power*p_prior/(p_alpha*(1-p_prior)+p_power*p_prior)); ## the target percentage should be 99.91% rather than 0.99% ##If you guess the point has 22% probability to fall in the H_1 region, you will update the probability to 82% after you know it fall in the light gray region with a significant exp report.

##to demo Prosecutor's fallacy on wiki
p_prior=10e-7;##if the defendant is not different From any one of the 10 million population except the DNA test result
p_alpha=0.0001;##"it is testified that the probability that two DNA profiles match by chance is only 1 in 10,000"
p_power=1; ##Let's assume the power is extreme
##
(h_1_s=p_power*p_prior/(p_alpha*(1-p_prior)+p_power*p_prior));
## the Procecutor mistook the 0.99% target percent as 99.99%.

##-------------------------------

##to demo the Defendant's fallacy
p_prior=10e-2;##"The video (evidence) suggests a 99% (rather than 1-10e-7) chance that the defendant is innocent", except the DNA test result
p_alpha=0.0001;##as the former
p_power=1; ##as the former
##
(h_1_s=p_power*p_prior/(p_alpha*(1-p_prior)+p_power*p_prior));
## the target percentage should be 99.91% rather than 0.99%

]]> By: iris http://lixiaoxu.lxxm.com/classic-neyman-pearson-approach-demo/comment-page-1/#comment-73 iris Sat, 01 Dec 2007 16:23:50 +0000 http://lixiaoxu.lxxm.com/2007/11/30/classic-neyman-pearson-approach-demo/#comment-73 Easy to understand the R script ,but hard to understand the meaning... Easy to understand the R script ,but hard to understand the meaning...

]]> By: lixiaoxu http://lixiaoxu.lxxm.com/classic-neyman-pearson-approach-demo/comment-page-1/#comment-72 lixiaoxu Fri, 30 Nov 2007 11:04:40 +0000 http://lixiaoxu.lxxm.com/2007/11/30/classic-neyman-pearson-approach-demo/#comment-72 It is interesting to verify whether two independent significant repetitions will persuade peers more than the combined whole data set does. It is interesting to verify whether two independent significant repetitions will persuade peers more than the combined whole data set does.

]]>