样本相关系数是总体相关系数的无偏估计量吗?


14

这是真的,是一个无偏估计ρ X ,ÿ?也就是说,ë [ - [R X ,ÿ ] = ρ X ,ÿ?RX,YρX,Y

E[RX,Y]=ρX,Y?

如果没有,什么是一个无偏估计?(也许有一个标准的无偏估计器被使用?而且,它类似于无偏样本方差,我们可以简单地进行调整,将有偏样本方差乘以nρX,Y?)nn−1

人口相关系数被定义为同时将样品相关系数被定义为- [RX,ÿ=Σ Ñ 我= 1(X我- ˉ X)(Ý我- ˉ ÿ)

ρX,Y=E[(X−μX)(Y−μY)]E[(X−μX)2]E[(Y−μY)2],
RX,Y=∑i=1n(Xi−X¯)(Yi−Y¯)∑i=1n(Xi−X¯)2∑i=1n(Yi−Y¯)2.

关于ρ的估计量的一个(有点类似)问题。ρ
— ttnphns

“什么是无偏估计量”问题以存在一个且只有一个为前提。 先验的,似乎没有任何理由认为。
— Michael Hardy

@MichaelHardy:我已经纠正了。感谢您指出。
— 肯尼·路易斯

刚刚偶然发现了这个线程,我认为这可能是一个有趣的阅读sciencedirect.com/science/article/pii/S0167715298000352(我自己尚未读过,
— TBH

Answers:


10

这不是一个简单的问题,但是有些表达式可用。如果您特别在谈论正态分布,那么答案是否定的!我们有

Ëρ^=ρ[1个-(1个-ρ2)2ñ+Ø(1个ñ2)]

ñ-2

ρ=0|ρ|=1个1个ñ

Ëρ^→ρ。注意,这已经通过有界收敛定理从样本相关系数的有界性和一致性中得出。


2
There may be infinitely many terms in the expression above, but "infinite terms" would be there are some terms, each of which is infinite.
— Michael Hardy

Suppose all points in a bivariate population lie on a straight line with nonzero slope. Then all points in any sample do so too. I conjecture therefore that if population correlation has absolute value |ρ|=1 so also sample correlation |r|≡1.
— Nick Cox

@NickCox That's true, in the degenerate case the sample correlation coefficient would return |1| with no estimation error.
— JohnK

For a related question, does anyone know if analogous results exist for any other distributions besides the 2D normal?
— Riemann1337
By using our site, you acknowledge that you have read and understand our Cookie Policy and Privacy Policy.
Licensed under cc by-sa 3.0 with attribution required.