是有些误会,

来源: 2013-03-13 13:55:04 [旧帖] [给我悄悄话] 本文已被阅读:

empirical cdf的最大值也只能是1,所以variance还是0。你给的link是说empirical cdf的variance。Wiki也有定义,http://en.wikipedia.org/wiki/Empirical_distribution_function#Definition

看看这个对你有用吗:https://www.scss.tcd.ie/Rozenn.Dahyot/ST2002/20080130IntroductionSE.pdf

这里说的那个Dirac delta function,就不是我能理解的了。就如wiki说的,“From a purely mathematical viewpoint, the Dirac delta is not strictly a function, because any extended-real function that is equal to zero everywhere but a single point must have total integral zero。”, 可是这个function的integral = 1。估计measure theory好的人才能看明白你的问题了。