<IF> Bootstrap was reliable
- From: "\"Luis A. Afonso\"" <licas_@xxxxxxxxxxx>
- Date: Wed, 28 Dec 2005 16:14:54 EST
If Bootstrap was reliable, it must provide
the same results than Student´s in the case
of normal samples. (Enjoy , please)
_X=N(115,20):20__Y=N(100,20):20
Diff. sample means____CI (Student)
____________________Bootstrap
1____25.28__________12.9___37.6
____________________8.9___ 42.4
_____________________________
2.___14.24___________2.6___25.9
____________________0.5___28.2
_____________________________
3___ 27.14__________15.5___38.8
___________________15.1___40.0
_____________________________
4.___12.76___________0.2___25.3
___________________-6.4___31.9
_____________________________
5___15.63___________1.5___29.8
___________________-2.0___33.1
_____________________________
6___ 26.06__________15.0___37.1
___________________12.9___37.9
_____________________________
7___20.65___________9.9___31.4
___________________ 4.8___35.8
_____________________________
8___10.19__________-5.8___26.2
__________________ -8.9___30.5
_____________________________
9___9.03___________-3.6___21.7
_________________ -10.4___28.1
_____________________________
10 _15.91___________3.2___28.6
__________________-2.5___ 33.6
_____________________________
The bootstrap C.I. is wider than those
provided by the Student’s algorithm.
Note that similar experiments, this time
With sample of size 30 one have the
same result.
_____licas (Luis A. Afonso)
.
- Prev by Date: Re: Maximum likelihood estimator and multiple maxima
- Next by Date: Re: Easy correlation questions
- Previous by thread: Easy correlation questions
- Next by thread: Problem on Balls and Urns
- Index(es):