Documente Academic
Documente Profesional
Documente Cultură
25 al 27 de Octubre
1.
INTRODUCCIN
25 al 27 de Octubre
x=
1 n
xi ,
n i =1
W=
s X2 =
1
n
( xi x ) 2 .
i =1
n 1
X X
T=
x2 = X2 / n ,
sx2 = s X2 / n .
Tambien
podemos
tomar
como
x = x2
un
x = ( x1 + x2
X X
. (2)
SX
sx2 , o sx , podemos
( x t( n 1,1 / 2 ) sx , x + t( n 1,1 / 2) sx ) ,
(3)
donde
dado
de
la
estimacin
+ ... + xn ) / n , tenemos solo n 1
xi
2
X
2
X
valor
observaciones
Z
=
W
(n 1)
(1)
varianza
(n 1) S X2
25 al 27 de Octubre
ef . =
Y es resultado
X 1 , X 2 ,..., X k ,
T=
1 , 2 ,..., k
sus grados
y Y
=
Sy
y Y
2
Y
2
Y
Y2 :
Z
.
W
(6)
ef
Y = f ( X , X ,..., X ) ,
1
f
Y2 = i =1 X2 i .
xi
expandida
y = f ( x1 , x2 ,..., xk ) ,
combinacin
2
1
2
2
lineal
de
las
incertidumbres
2
k
f
sY2 = i =1
s X2 i .
i ( x1 , x2 ,..., xk )
k
influencia.
sY2 ,
( y t( ef . ,1 / 2 ) s y , y + t( ef . ,1 / 2) s y ) ,
3.
(4)
y es el resultado de medicin, s y su
t( ef . ,1 / 2)
tiene
combinada,
incertidumbre
donde
(5)
de libertad correspondientes.
si4
i =1
para
s y4
es
APROXIMACIN DE WELCHSATTERTHWAITE
Sean
W1 ,W2 ,...,Wk ,
variables
aleatorias
el
25 al 27 de Octubre
con
r = r1 + r2 + ... + rk
grados
de
E (U ) =
V (U ) = E (U ) 2 [ E (U ) ] = 2 .
2
U
E ( ) = 1,
U
2
E ( ) 2 = 1.
E ( i =1 aiWi ) = 1 , y
k
E ( i =1 aiWi ) 2 =
k
aiWi y
i =1
aiWi
(9)
1.
(10)
i =1
(8)
k
U
E ( i =1 aiWi ) 2 = E ( ) 2 .
(7)
k
U
E ( i =1 aiWi ) = E ( ) ,
2
E ( i =1 aiWi ) 2 1
k
(11)
25 al 27 de Octubre
2
( i =1 aiWi ) 2 1
k
E ( k aiWi )
i =1
.
=
2
k a
i=1 i E (Wi )
k
2
V ( i =1 aiWi )
k
= E ( i =1 aiWi )
+ 1 ,
2
k
E ( aiWi )
i =1
entonces
V ( i =1 aiWi )
k
E ( i =1 aiWi ) =
k
E ( k aiWi )
i =1
+ 1.
(12)
distribuidas
2 i =1
k
ai2 [ E (Wi ) ]
Wi
(14)
ai = (f / xi ) , y Wi = si2 .
i =1
ai2
2 E ( i =1 aiWi )
.
=
k
V ( i =1 aiWi )
E ( i =1 aiWi ) = 1 ,
( i =1 aiWi ) 2
(13)
V (Wi ) = 2 [ E (Wi ) ] / i .
2
RESTRICCIONES O LIMITACIONES
25 al 27 de Octubre
REFERENCIAS
[1] Guide to the Expression of Uncertainty in
Measurement,
Geneva,
International
Organization for standardization, 1993, ISBN
92-67-10188-9.
[2] Casella, G. and Berger, R. L., Statistical
Inference, Wadsworth & Brooks/Cole, Pacific
Grove, CA. 1990.
[3] Welch, B. L., The Specification of Rules for
Rejection too Variable a Product, with
Particular Reference to an Electric Lamp
Problem, Supplement to the Journal of the
Royal Statistical Society, 3(2), 1936, 29-48.
[4] Welch, B. L., The Significance of the Difference
Between two Means when the Population
Variances are Unequal, 29, 1938, 350-362.
Satterthwaite,F. E., Psychometrika, 1941, 6(5),
309-316.
[5] Satterthwaite,F. E., Psychometrika, 1941, 6(5),
309-316.
[6] Satterthwaite,F. E., An approximate distributin
of
estimates
of
variance
components,
Biometrics, Bull., 2(6), 1946, 110-114.
[7] Welch, B. L., The generalization of Students
problem when several different population
variances are involved, Biometrika 34, (1947),
28-35.
[8] Kendall, M. and Stuart, A., The Advanced
Theory of Statistics, , Vol. II: Inference and
CONCLUSIONES
25 al 27 de Octubre