Forgot password
 Register account
View 2236|Reply 3

[不等式] 请教一个指数不等式题

[Copy link]

45

Threads

51

Posts

0

Reputation

Show all posts

等待hxh posted 2014-12-10 21:39 |Read mode
指对数不等式.png

673

Threads

110K

Posts

218

Reputation

Show all posts

kuing posted 2014-12-10 22:29

24

Threads

1017

Posts

46

Reputation

Show all posts

战巡 posted 2014-12-11 06:58
Last edited by 战巡 2014-12-11 07:10回复 1# 等待hxh


开发一种新方法:

这个等价于证明:
\[\sum_{i=0}^n\frac{n^i}{i!}e^{-n}>\frac{1}{2}\]
\[P(X\le n)>\frac{1}{2}, X\sim POI(n)\]
而对于$P(X\le m)=\frac{1}{2}$的这个$m$,显然就是泊松分布的中位数
对于泊松分布,可知其偏度(Skewness)为:
\[E[(\frac{X-\mu}{\sigma})^3]=E[(\frac{X-n}{\sqrt{n}})^3]=\sum_{i=0}^\infty (\frac{i-n}{\sqrt{n}})^3\frac{n^i}{i!}e^{-n}=\frac{1}{\sqrt{n}}>0\]
因此泊松分布为正偏分布,而且由于泊松是单峰分布,就有其均值$n$是大于中位数$m$,因此有
\[P(X\le n)>P(X\le m)=\frac{1}{2}\]

又显然当$n\to \infty$时,有偏度趋于0,也就是中位数趋于均值
因此
\[\lim_{n\to \infty}P(X\le n)=\frac{1}{2}\]

673

Threads

110K

Posts

218

Reputation

Show all posts

kuing posted 2014-12-11 14:28
高科技,只能潜了

Quick Reply

Advanced Mode
B Color Image Link Quote Code Smilies
You have to log in before you can reply Login | Register account

$\LaTeX$ formula tutorial

Mobile version

2025-7-21 12:07 GMT+8

Powered by Discuz!

Processed in 0.016426 seconds, 28 queries