--
You received this message because you are subscribed to the Google Groups "ZHANGroup-seminar" group.
To post to this group, send email to zhangrou...@googlegroups.com.
To unsubscribe from this group, send email to zhangroup-seminar+unsub...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/zhangroup-seminar?hl=en.
我是这样想的。
抛一枚硬币n次,m次朝上的概率(p1)为服从二项分布,即:p1 = C(n,m) * (0.5)^m * (0.5)^n。
但是作为抛硬币的主体,人可能会看错,使观察值不等于实际值。
问题演变成,抛一枚硬币n次,在肉眼可能出错的情况下,观察到有m次朝上的概率(p2)是多少?
记一次观察中,把正面看成反面的概率为p(+/-),把反面看成正面的概率为p(-/+)
那么,抛一次硬币,
实际值为正,观察值为正的概率为 0.5*(1- p(+/-));
实际值为正,观察值为反的概率为0.5* p(+/-);
实际值为反,观察值为正的概率为0.5*p(-/+);
实际值为反,观察值为反的概率为0.5*(1- p(-/+));
这样,抛一次硬币,
观察值为正的概率为0.5 * (1 - p(+/-) + p(-/+));
观察值为负的概率为0.5 * (1 – p(-/+) + p(+/-))
故:p2 = C(n,m) [0.5 * (1 - p(+/-) + p(-/+))]^m * [0.5 * (1 – p(-/+) + p(+/-)]^n
一般可以认为p(+/-) = p(-/+);那么 p1 = p2
To unsubscribe from this group, send email to zhangroup-semi...@googlegroups.com.