Statisticians_group : Message: statistics

Sign In
New User? Register

Statisticians_group

Yahoo! Groups Tips

Did you know...

You can set the sort order of messages? Just click on the link in the date column. Your preferences will be remembered, so you don't have to do it again when you return.

Messages

Messages Help

Advanced

statistics

Message List

Reply | Forward

Message #1217 of 4072 |

Hi everybody
What better estimations for P(|X-EX|>t) do we have instead of
Tcebyshev's one?
Gh

Sun Apr 2, 2006 5:14 am

Show Message Option

"gheorghitazbaganu" <gheorghitazbaganu@...>

gheorghitazb...

Offline

Send Email

Forward

Message #1217 of 4072 |

Expand Messages	Author	Sort by Date
Hi everybody What better estimations for P(\|X-EX\|>t) do we have instead of Tcebyshev's one? Gh...	gheorghitazbaganu gheorghitazb...	Apr 2, 2006 1:50 pm
Hi, This Chebychev's (or Tcebyshev's... spelling varies from book to book; not sure what is the right) inequality P(\|X-EX\|>t)<=var(X)/t^2 holds for all X which...	Madan Gopal Kundu madan4331	Apr 2, 2006 7:40 pm
Some other inequalities you might be interested in are Gauss inequality (that assumes a unimodal distribution), or the following MGF based inequality: Prob[ X...	Stanislav Kolenikov skolenik	Apr 4, 2006 2:06 am
Yes, it depends on the tails. If F is short tailed, and a is great, i agree that the mgf inequality is better than Cebychev's one. But in general? I do not...	Zbaganu Gheorghita gheorghitazb...	Apr 4, 2006 8:22 pm

< Prev Topic | Next Topic >

Advanced