統計問題

2 views
Skip to first unread message

Su小欣

unread,
Mar 28, 2008, 9:54:57 PM3/28/08
to 生態演化討論小組
最近在準備統計學考試時,看到書上寫著一個基本原理:

When sampling from any distribution with finite variance, the sample
mean and sample variance are unbiased estimators of population mean
and variance, respectively.

雖然這文句我有做了一些調整,而且我不知道該如果在這網頁上打出數學符號(X-bar, sigma-squared...),但我想問題的是,有什麼
distribution會是沒有finite variance的嗎?那會是用來描述什麼情形(如何推導出來的?)?

謝啦!

Yi-Huei 怡慧

unread,
Mar 29, 2008, 10:19:09 AM3/29/08
to 生態演化討論小組
小欣,
你一定是看數理統計的書
finite才能保証"存在",這在証明一些比較數學東西時要用到.

Su小欣

unread,
Apr 1, 2008, 8:30:38 AM4/1/08
to 生態演化討論小組
是啊,目前在修高統二,用的是Hogg的 Probability and Statistical Inference課本,雖然我覺得尚且稱不上
是"數理統計",但對我而言的難度也"夠了"! ToT

原來"finite才能保證存在"哪......所以,基本上統計學內是根本沒有什麼非finite variance的distribution囉~?

Yi-Huei 怡慧

unread,
Apr 2, 2008, 6:46:12 AM4/2/08
to 生態演化討論小組
有呀!
Cauchy distribution的mean and variance do not exist!
我不曉得你的書有沒有這分佈.Casella那本的附錄有這個分佈.

還有Geometric distribution的variance 是 (1-p)/(p^2),as p near 0, this
variance will not exist.
> > On Mar 29, 9:54 am, Su小欣 <sush0...@gmail.com> wrote:- Hide quoted text -
>
> - Show quoted text -

Su小欣

unread,
Apr 3, 2008, 12:16:38 PM4/3/08
to 生態演化討論小組
喔! 對啊! 我們有教一點點Cauchy,好像有聽老師提到過。倒是Geometric,我都沒想過當p趨近於0會怎樣....

謝謝啦~
> > - Show quoted text -- 隱藏被引用文字 -
>
> - 顯示被引用文字 -
Reply all
Reply to author
Forward
0 new messages